This post was rejected for the following reason(s):
This is an automated rejection. No LLM generated, heavily assisted/co-written, or otherwise reliant work. An LLM-detection service flagged your post as >50% likely to be written by an LLM. We've been having a wave of LLM written or co-written work that doesn't meet our quality standards. LessWrong has fairly specific standards, and your first LessWrong post is sort of like the application to a college. It should be optimized for demonstrating that you can think clearly without AI assistance.
So, we reject all LLM generated posts from new users. We also reject work that falls into some categories that are difficult to evaluate that typically turn out to not make much sense, which LLMs frequently steer people toward.*
"English is my second language, I'm using this to translate"
If English is your second language and you were using LLMs to help you translate, try writing the post yourself in your native language and using a different (preferably non-LLM) translation software to translate it directly.
"What if I think this was a mistake?"
For users who get flagged as potentially LLM but think it was a mistake, if all 3 of the following criteria are true, you can message us on Intercom or at team@lesswrong.com and ask for reconsideration.
you wrote this yourself (not using LLMs to help you write it)
you did not chat extensively with LLMs to help you generate the ideas. (using it briefly the way you'd use a search engine is fine. But, if you're treating it more like a coauthor or test subject, we will not reconsider your post)
your post is not about AI consciousness/recursion/emergence, or novel interpretations of physics.
If any of those are false, sorry, we will not accept your post.
* (examples of work we don't evaluate because it's too time costly: case studies of LLM sentience, emergence, recursion, novel physics interpretations, or AI alignment strategies that you developed in tandem with an AI coauthor – AIs may seem quite smart but they aren't actually a good judge of the quality of novel ideas.)
Torture of Prometheus (circa 1646 to 1648) Salvator Rosa
My first post here, feedback greatly appreciated.
Epistemic status: Reflective essay / argument sketch about the epistemic value of fiction in the LLM era. Confidence low-to-moderate; many claims are taste-level and philosophical rather than empirical. Seeking counterexamples, relevant studies, and pointers to prior LW discussions. Epistemic effort: Several close readings (Dostoyevsky, Tanizaki, Locke) + familiarity with public statements by Sutskever/LeCun; no new data.
More recently, I’ve been preoccupied with why we still read novels. With a wealth of entertainment opportunities from movies to short form videos, a book can feel like large time commitment, and in a distracted world can feel more like a burden than enjoyment.
There is also a plethora of pretentious reasons to read a book. Certain ones are prestigious and can falsely elevate our perceived self-importance. They can also be tick box exercises, especially when received as a gift or loan.
On a personal note, this is the feeling I had when beginning ‘The Brothers Karamazov’, a book touted as the epitome of the form. As a 1000-page marathon, the prestigious charm and daunting compulsion to enjoy every word quickly fades. Yet still, it was completely transfixing, far exceeding my expectations.
Beyond the personal, classics help shape our collective understanding. Era defining books (novels or otherwise), demonstrate or introduce a new idea, concept, or aesthetic feeling. These ideas often leach out into our sociocultural structures, spreading their roots into other disciplines. They sometimes spawn entire trees of literary work, and looking back can feel like tracing the phylogenetic tree of human ideas. Great classics ride the fine line of being tangible enough to be compelling, but too abstract to be yet equation or scientific theory. For Tanizaki, it’s the aesthetic importance of negation and shadow1, an idea already present in Japanese culture implicitly, and for Locke it’s the futility of using words to represent reality2 (ironic as he writes in essay form).
But still, not all classics are novels. Beyond entertainment, how fruitful is fiction? There’s a qualitative difference between novels in comparison to scientific papers or philosophical essays. Novels show us what may otherwise be told to us. Nietzsche’s philosophical essays tell us about the nihilistic abyss, like a friend recounting a place he has been. In contrast, Dostoyevsky expounds his philosophy through the internal dialogues of his masterfully crafted characters and their actions. He pushes us into the abyss, drowns us in its experience, before dragging us out by our hair, enriched and nourished.
The subjectivity inherent in fiction can make it frustrating, but is sometimes a blessing. When sufficiently obscure, novels lay down a conceptual framework for us to explore and formulate our own ideas. We generate the world procedurally according to the authors recipe, and there’s something more enriching about preparing a meal yourself.
But is that knowledge unshakeable? Is great literature a stone in the edifice of human knowledge, or a painting overlayed?
To zoom in, words convey meaning in relation to the representations they conjure in others minds. They are a reliable output filter for a chaotic vortex of internal representations, made communicably rigid, but dynamic enough to evolve alongside culture and ideas. Still, to be ‘articulate’ as a storyteller, there is a over-reliance on the ability of an individual to anticipate the imprints that his words will make on others. How else can he select the ‘best’ form. Unchecked, this can develop into pretentiousness and superficiality, and I’m increasing skeptical of literary work that gets the ratio of ideas to emotional impressions wrong3.
A limitation of storytelling formats is that concepts can be restricted and bound by this hyper-awareness of the audience. The best writers are cognizant of this weakness. Dostoyevsky seems to be particularly scathing towards the literary type in ‘White Nights’4. His protagonist can only conceive of events as emotional matter to be reworked, intensified and replayed in the manner of a Weltschmerzen5.
Still, reading carefully with good intentions can bear impressive fruits. Some literary complexity, far from being pretentious, is an attempt to reason through a very complex sentiment or feeling using well crafted analogy. Like attempting to schematise and draw the outline of some idea. It instils an appreciation and wariness of how compelling narratives help appropriate profoundly complex phenomena into worldviews digestible to a profoundly limited analogue intelligence.
As for digital intelligence, these questions are more vital after the arrival of Large Language Models. By feeding all published human corpus into complex neural networks capable of imperceptibly fine computations, silicon too has mastered language. With a few prompts, we can elicit text in any given form or style about almost any topic. There is clear danger in unrestricted to access to fine tuned persuasive language.
But can LLM’s do much more than perform a stylistic and seductive dance? Are there abstract representations that will persist and contribute towards science. Ilya Sutskever maintains that an LLM’s abstractions hold weight in that its training has a compressed representation of all written human ideas6. Conversely, Yann Lecun argues that with regards to intelligence they barely compare to a cat, in that they the lack the wealth of information omitted by language7.
An uncontested fact is that AI has democratised, to an extent, the art of being articulate. But how will language evolve in this age? The advent of LLM’s signifies an event horizon with regards to how we consider anything literary. As to the predictions of the ‘Underground Man’8, perhaps humanity will elevate text with imperfection and idiosyncrasy simply to undermine the finely tuned, mechanistic sentences of AI, preserving our fragile egos.
In such an age, the human writer will transition to the equivalent of a traditional craftsman. His work will be credited not for its utility, but for its aesthetic appeal. The handcrafted object is a source of deep satisfaction, but remains a mere novelty. AI has insulted the perceived superiority of human language, and many more intellectual insults are to come. But still, it cannot be devalued. The dynamism and breadth of literary mastery continues to render it a vital toolset, especially in navigating the gulf of existential angst that lies ahead.
References:
1. Jun’ichirō Tanizaki, In Praise of Shadows
2. John Locke, Of the Abuse of Words
3. Friedrich Nietzsche tilts towards emotion. Paul Graham sits at the opposite end of the spectrum.
5. ‘Weltschmerz (German: literally “world-pain”) is a literary concept describing the feeling experienced by an individual who believes that reality can never satisfy the expectations off the mind, resulting in a “mood of weariness or sadness about life arising from the acute awareness of evil and suffering”. From https://en.wikipedia.org/wiki/Weltschmerz
