Robert Nozick (in Examined Life) asked how we feel if we found out, say, Beethoven seriously composed music based on a secret formula, which is entire mechanical and required no effort for him at all.
Would we still appreciate the music in the same way? If not, does our appreciation really stem from the fact that we feel he has also struggled like we do, and nevertheless produced something incredible.
I remember as a very small child watching figure skaters on TV and thinking "that's no big deal". And before I started programming: "it's just logic, all very straightforward". But that was before I first entered an ice rink or centre-d a div
Maybe we don't really appreciate something unless we appreciate it is hard in a visceral way.
> Beethoven seriously composed music based on a secret formula, which is entire mechanical and required no effort for him at all.
If he discovered the formula, then yes I imagine most people would appreciate the music just as much, if not more so.
If he copied the formula from somebody else, then he was just turning the crank - a far more sterile and mechanical affair.
Using Suno to "create" music is just turning the crank.
Related but much of Bach's music is just as incredible for its incredible mathematically relational structures as it is for its pure virtuosity and brilliance.
To clarify, what Nozick meant was what if Beethoven was just turning the crank.
"Yet our experience of Beethoven's string quartets would be diminished if we discovered he had stumbled upon someone else's rules for musical composition, which he applied mechanically". (p. 38 of https://archive.org/details/examinedlife00robe/page/n15/mode...)
I guess another way to put the question is this. Suppose there is an alien civilisation where their brains are hard-wired to make Beethoven level music automatically. Most of us can hum a tune without effort: these aliens can hum music that would strike us as original and compelling without much effort. How would we react then?
Plus: nice pointer on the math scholar link. I remember loving the musical parts of Godel Escher Bach. Wish there is a good interactive website where I can revisit all the content (and listen to all the music) there in the browser.
This stolen valor mindset becomes absurd if Bach anonymously open-sources the formula he discovered and dies before releasing any music himself.
Whatever music that follows, however hard fought for - even if 1:1 Bach output that he kept in a drawer, isn't beautiful? Just string plucking? That's not music appreciation, that's love of reputation you can easily grasp and associate with.
I think Nozick's example is meant to make us re-think whether there is a strict separation between music (or other artistic) reputation and love of reputation.
In Anarchy, State and Utopia, he tackles some utopian theorists' claim that, if equality prevails, everyone will rise up to the level of the greatest writers and artists. Would people be content then? Or will they still want to vie for "eyeballs"? If the latter, should we just admit that there is just a deep-seated human desire to compete for dominance?
The artistic value isn't magically generated by the piano itself, nor by the LLM in isolation. It's the result of the skilled interaction, the human artistry expressed through this new and powerful instrument.
> Maybe we don't really appreciate something unless we appreciate it is hard in a visceral way.
Count me out of that "we" -- I appreciate the artist who put in the work because they put in the work to make the thing I like, but I don't appreciate the thing because of the effort. I can marvel at the effort required to produce art in a certain way, but I'm (largely) indifferent to the effort in my actual appreciation of the thing (or lack of it).
I look forward to the time when I can have as much high-quality (to me) fiction to read as I like, because it's all generated by LLM. Some time after that, I'd love to see the main Star Wars sequence done properly. I won't care that it isn't created by a vast team of humans.
To represent the other side, I enjoy reading Urdu and Persian poetry but I will never be interested in reading anything generated by an LLM. No matter how 'high quality' it is represented to be, I'm aware that it was produced by a process that shares nothing with my own experience of the world. It has felt no hope, disappointment, fear, pain, mortality, loss of loved ones, lack of control over itself, a world model that has changed over the years, and doesn't know that it all doesn't amount to much in the end and yet this is all there is for itself. It may turn out to be sentient in some way, but it's almost certainly not sentient in the way that I am sentient. I know it's just mimicking being human as instructed; to take it seriously devalues everything about my own humanity. I'm not ready for that kind of enlightened insight, I think.
> I know it's just mimicking being human as instructed; to take it seriously devalues everything about my own humanity.
Since it is mirroring human culture, why do you see it in such a negative light? Instead see it like what it is, an interactive reconstruction, or maybe like a microscope to zoom into any idea.
I’m happy to use LLMs in all other contexts, quite enthusiastically actually. I’ve got DeepSeek 32B running locally on a beefy PC already.
It’s just in the context of poetry, and literary writing in general, that I feel differently about them. There’s also the fact that I haven’t read all that human writers and poets have already written (and will never be able to in this short life) so there’s no need to turn to synthetic output. No supply problem exists. Poetry in particular is something to ponder over and over. You can’t really run out.
> It has felt no hope, disappointment, fear, pain, mortality, loss of loved ones, lack of control over itself, a world model that has changed over the years, and doesn't know that it all doesn't amount to much in the end and yet this is all there is for itself.
You can't know what any poet felt or didn't feel while writing a poem. Perhaps it was a commission piece, or an experiment or an emulation of something the poet had heard elsewhere.
And more generally, whether the specific emotion another man feels is similar or even comparable to your own is also unknowable. He might use the same word to describe it, but the subjective experience associated with it might be completely different, and completely impossible to share.
Yes but at least it was possible for that poet to have felt what I feel they might have felt while writing that poem. And the closer they are to me culturally the more likely it is that I am not misidentifying their emotions entirely.
Also poems are not really puzzles to be solved. If it produces an effect and is solid craft-wise, that is enough. There’s a lot to the craft side btw in the Urdu and Persian ghazal form which is what I had in mind while writing my original comment. LLMs can easily master the latter but have nothing to do with the former. Their output is pure form without substance.
Edit: I want to add that ambiguity (ابہام) is even a desirable property in Urdu ghazal, specifically. The more interpretations a couplet can have, the greater is the accomplishment in terms of craft.
I think it's possible to slice preferences endlessly, and I do think I'm a bit unusual.
I gave ChatGPT a list of my favorite SF novels, and a brief description of why, and asked for similar works. It recommended 10 novels, three of which I've read and weren't in the sweet spot. Also, everything it recommended was 30+ years old -- to be fair, the same is true of the list I gave it, but I think it goes against your point that there's an unlimited supply.
So I told it about the three and asked it to adjust and to give more recent works, and it obliged. One of the new recommendations was in the Culture series, which I've read one of and it wasn't my jam. Another was Project Hail Mary by Andy Weir, which I've read and enjoyed the Martian, but I'm betting that's the only Andy Weir I'll like. The others I'll have to check out.
> I look forward to the time when I can have as much high-quality (to me) fiction to read as I like, because it's all generated by LLM. Some time after that, I'd love to see the main Star Wars sequence done properly. I won't care that it isn't created by a vast team of humans.
I think the problem here is analogous to the "500 channels and nothing to watch" issue in the heydey of cable.
Ok let's say you have an LLM in your hand that can generate any story you want. High quality. So you say: "tell me a story" and it tells you as story. But what story? Who is in it? What characters? Why are they there?
The only novelty that's going into this the prompt. Everything else is regurgitated weights and probability associations. The question is: does the full infinite closure of recombination over some finite learning set (no matter how large) encompass enough of the essence of creativity to produce something "new"?
This is a hard question to answer because it forces us to try to define creativity, or lacking that - at least try to identify where it comes from.
I don't have a clear answer to this but I'll suggest a line of thinking that seems plausible.
When a person writes a story, it's not derived as an amalgam of everything they have read. It's not some probabilistic weighted average of all those associations. The story they write is also derived from their lived experience. Their personal interactions, their observations, their musings, their passions, their fears.. and how all of those things interact with their circumstance, influencing their reactions, those reactions influencing their environment, and that feeding back into the above process.
There are two components that seem important here: the first is the existence of a rich, dynamic, and active _dialogue_ between the mind and its environment. It's not static, and it involves a feedback loop between the mind and the environment it models.
The second is a motive force. For humans the origin is biological. Fear, hunger, satiation, arousal, etc. - those core primitive emotional drives that originally developed to help us survive, but then were layered over with an intellect that elaborated on them. What originated as a motive force to drive the mating instinct evolves into a sonnet about an unattainable maiden. The fear of the dark that keeps us away from the places where we would be eaten.. evolves into a stories about unfathomable creatures and impossible colours that drive men insane.
And I think there's a third one that's unelaborated and implied but should be made explicit: introspection & reflection. The ability to consider your choices and consequences with respect to your motivations, and adjust any number of things - from the motivations themselves, to expectations/understanding.
This creature would have a lived experience, some underlying motivations, and a feedback loop established between the two using introspection. I have no idea how you'd build any of that.. but it feels like that's what you'd need before you got yourself a good storyteller.
But by that point, you'd also be compelled to question whether or not it's even ethical to force it to tell you a story anymore.
I don't think it's impossible that some broader AI system eventually is capable of genuinely creating creative output. LLMs are not that, though.
The value of art is in meaning and context.
Purely generative art is as meaningful as a pretty rock.
Think of the models as a camera.
If you take shots from a car's dash cam in a city at random you might fall upon some really beautiful photos. But this is chance, the camera didn't create the city or its scenes.
A photographer can choose or create meaningful scenes because he has a mind, consciousness and life experience.
Pretty rocks (e.g. mountains, gems, etc.) are frequently ascribed substantial meaning, despite the fact that no consciousness had a hand in creating them.
Raw gemstones are generally uninteresting until people shape them. Diamonds worth thousands may not even qualify as interesting enough to pickup in the raw state, assuming you don’t know what it is.
Mountains get meaning as aspects of our environment, but try and name the your top 10 most aesthetically pleasing mountains. At least for me, I may appreciate a scenic view but I just don’t think of them in that kind of context.
> A photographer can choose or create meaningful scenes because he has a mind, consciousness and life experience.
But so does a user. Users don't prompt "draw a dog" but give 3 lines of intricate details and iterate a dozen times until it looks right. It's not like these models work all on their own.
There are people who define art that way, and there are those who define it as beautiful things. I'd personally own and display something beautiful made by algorithm than much post modern art, which is frequently visually unpleasant in spite of being rich in some message.
Our brains are drawn to some things visually for instinctive reasons, and I don't need a big message when I'm decorating or wanting to please the eye.
That's a very reductive and limited way to look at art. You're right in stating it as decoration but I would not conflate the two. Different things, both valuable in its own right.
A rose is beautiful, a painting of a rose less so, its two dimensional and lacks many aspects of the true rose. But it has more value because there are millions of roses, while the painting captures a unique experience of viewing the rose by the painter and transmits it to the other person viewing the painting.
Unless there are ghosts in the shell, MidJourney gives an aproximation of how a painting of a rose looks like. Its like an aggrregate function that averages a million artists.
Its a weird concept.
So is it still art if a photographer is driving the car with a dash cam and they drive it with the intention of capturing great images, and then goes over all the captured frames to find the best ones?
I would say yes, this dash cam technique can be an artistic method. Reminds me of Jon Rafman's wonderful Nine Eyes project - he captures screenshots from Google Streetview, see
https://9-eyes.com
You’re assuming there is any coherent or consistent epistemological grounding for the average person’s beliefs
Its a fools errand - it’s an infinitely small set of people who can accurately describe their reasoning - even fewer have consistent reasoning - fewer still have coherence between beliefs
The ones that do we call either monks or crazy
I’d argue people aren’t even coherent enough to know how or what to appreciate
And, of course, you happen to be one of them no doubt :)
The classic internet philosopher’s lament: everyone else is irrational, inconsistent, and incapable of coherent thought—except, of course, the enlightened commentator making the claim. The irony is that this kind of sweeping generalization is itself an incoherent mess, built on vague cynicism rather than any serious engagement with human reasoning. If you actually believe that consistency and coherence are so rare, what exactly do you think you’re demonstrating here? Because from where I’m sitting, it looks less like deep insight and more like self-important nihilism masquerading as wisdom.
> The classic internet philosopher’s lament: everyone else is irrational, inconsistent, and incapable of coherent thought—except, of course, the enlightened commentator making the claim.
But that wasn't the claim. @AndrewKemendo said "it’s an infinitely small set of people who can accurately describe their reasoning - even fewer have consistent reasoning - fewer still have coherence between beliefs" So he didn't say that everyone else is irrational, he said that very few can accurately describe their reasoning. And I think this is true. Very few take the time to introspect. Fewer still will do so to the point that they are consistent in their thinking. And fewer still will will analyze their values and beliefs and get them to square up with each other. Their is nothing controversial here. It's demonstratively true, all one has to do is listen to people carefully and probe them to motivate their reasoning every now and again.
> The irony is that this kind of sweeping generalization is itself an incoherent mess, built on vague cynicism rather than any serious engagement with human reasoning.
I reject that it's a "sweeping generalization" – I assert that most if not all people who spend enough time carefully introspecting and observing others necessarily must come to this conclusion. What about the claim is an "incoherent mess"? Clearly this is a personal peeve of yours because your response is emotional and doesn't refute the claim in any decent way.
> If you actually believe that consistency and coherence are so rare, what exactly do you think you’re demonstrating here?
That's a logical fallacy.
> Because from where I’m sitting, it looks less like deep insight and more like self-important nihilism masquerading as wisdom.
Have you considered that it isn't that these people don't understand or can't express their motivations, beliefs, and values, but rather that they feel zero need to justify them to you or anyone else who questions them with the sole intent of proving themselves correct?
Have you seen the documentary Tim's Vermeer? Its thesis is that Vermeer, through advanced lenses at the time, was able to paint essentially mechanically rather than having a fine grasp of artistic brushstrokes in the traditional sense. Some, as well as in the documentary itself, think that it'd ruin the purpose of the art but I see it differently, especially with all of these AI artists now online, where the intent of the human making the art is all that matters, not the instrument or manner in which they do it.
In my mind, art has always been a technological endeavor. Language, writing, and grammar are all tools. Brushes, stroke technique, and paint composition are all tools. I heard a story about Tony Hawk pioneering some skate board move, being the first in the world to get it right. And then seeing some teenagers doing the same thing years later in a park.
Real artists learn what is possible and then develop tools to break those limits.
The real value of almost everything is based on effort I think. The best gifts aren’t the ones that are the most expensive but the ones the giver put the most effort and time into. One of the reasons I like pre-cgi is the amount of skill and the effort put into FX is astonishing. Claymation and stop motion don’t look amazing - it’s the effort.
And to your point, knowing how much effort really goes into something often requires a bit of experience to really appreciate it.
The first two are stop motion, but not claymation. The third is claymation, but I'm hard pressed to call it "beautiful". Striking, to be sure, but also conspicuously ugly.
At this point Aardman is also doing a lot of non clay stop motion, but it's still the core of their work.
Appreciating a piece of music is purely on its own merits of its content not if it was easy to create or not.
The background, ethics, skill or even creative process of the people behind it have no bearing on whether the music itself is good and how we much we like it, even if Hitler wrote the 9th symphony it would be still be a just as good a masterpiece. To consider anything but the merits of the output is a slippery slope of what biases are acceptable and not, that inevitably ends up being racial or at least exclusionary.
Even it was not as difficult as you imagined it to be, he still was the first to find it, or even just the first to popularize it and that is all that matters.
I think it's fair to say that the context in which a work was created adds to the novelty and ingenuity of it's existence. These works don't exist in a vacuum and there's certainly a difference between a symphony created by Beethoven in his time/setting and a symphony produced as some model's output.
Sure they may functionally have the same effect or enjoyment, but appreciation of a fine work goes deeper than its function.
Robert Nozick (in Examined Life) asked how we feel if we found out, say, Beethoven seriously composed music based on a secret formula, which is entire mechanical and required no effort for him at all.
Would we still appreciate the music in the same way? If not, does our appreciation really stem from the fact that we feel he has also struggled like we do, and nevertheless produced something incredible.
I remember as a very small child watching figure skaters on TV and thinking "that's no big deal". And before I started programming: "it's just logic, all very straightforward". But that was before I first entered an ice rink or centre-d a div
Maybe we don't really appreciate something unless we appreciate it is hard in a visceral way.