Could you describe the movements relative to the rest of the scene you’re setting?
“Part of the mechanism moved along the west wall, as though guided by (mechanical/spooky action related to plot). It stopped, then rotated towards (character) but came to rest on its corner before completing even a quarter rotation to the east.”
My contribution to beehaw so far is this comment. I have confidence issues, and have been essentially radio silent on social media for the past ten years, but I’m at a turning point in my life and writing is the only thing that makes sense I need to start now. I have a few sketches of ideas but none fully realized.
I’m primarily interested in short stories which I would like to potentially narrate in the future, but I’m slowly trying to get over the confidence problem at the moment.
My contribution to beehaw so far is this comment. I have confidence issues, and have been essentially radio silent on social media for the past ten years
Well, I feel honored that you are responding to my post! 🙂
but I’m at a turning point in my life and writing is the only thing that makes sense I need to start now
Turning points are precisely where you change direction in your life, so it sounds like you are doing the right thing! 👍
I’m primarily interested in short stories which I would like to potentially narrate in the future, but I’m slowly trying to get over the confidence problem at the moment.
While I can’t possibly know exactly how you feel or what factors may have led to this, I do somewhat understand. I had general anxiety issues and maybe a tinge of depression when I put my feet on this path. I barely touched the project for the first 6 months or more after I decided I was doing it.
My really very amateur advice is - take positivity from everything you can. For example, what helped me was a BUNCH of honestly sort of small sounding things that seemed to bear fruit over time. Two examples -
I made a playlist called “Positive” and added any song to it that made me feel good. Even if it wasn’t explicitly a song about positivity - if it made me feel energized and happy to hear it, it went in. I would put that playlist on every time I sat down to write.
When someone crossed my twitter feed who looked like they were spreading positive feelings to people, I followed them to ensure I’d get those positive messages in my feed.
I’d also suggest finding a writing podcast that you enjoy listening to. I’ve been listening to a few of them - at least one episode a day, and on one of them I’m about to start their back catalog for the third time. There are a lot of them out there, and they are all just a bit different from each other. One or more will click with you. It will not only educate and motivate you, but it will also help you feel more confident. Every one of these published writers that I have listened to has complained about the same feelings of doubt regarding their writing, the same difficulties with confidence and impostor syndrome, and really almost all the same problems that we beginners seem to have.
Although I don’t intend to plug a particular one, the Writing Excuses podcast often discusses short fiction, and gives specific advice for short story writers. At least one of the regular hosts has extensive experience with short fiction. Their show is also a little bit different every season - they change up the structure of the show and have guest hosts, and other things to add variety and to cover a lot of facets of writing.
I’m not qualified to give you any advice about how to write, but I will say this advice from B. Dave Walters is in my head all the time - I’m paraphrasing: The worst thing you write is better than the best thing you don’t write. (because it’s the only way to get better and that won’t happen if it stays in your head)
True. I have to remind myself that that’s what’s so great about any art form; the audience’s unique interpretation of it. Everyone is going to see something differently, and each perspective is (in most cases) valid.
Just write to the best of your ability. Writing is a written medium, not a visual one, so unlike a film, TV series, or game, you can’t guarantee that a reader will always see exactly the same thing as you. The imagination is a funny thing, which runs a full spectrum from aphantasia (an inability to imagine at all) to fully photorealistic “movie in their head” mental images. Nothing you write will ever allow those on one end of that spectrum to see what you’re seeing, and writing all the details to make the other’s mental images 100% accurate will bog down your writing with a lot of detail that just isn’t relevant to the story. Also, the more specific you make your descriptions in an attempt to beam what you’re seeing into the readers’ heads, the denser your prose becomes, which risks losing readers with average reading comprehension skills.
If there’s a specific aspect of the movement that’s very important, then describe it as best as you can. Readers don’t need, and won’t remember, irrelevant details. If they do picture something incorrectly and something you write later contradicts it, then they’ll edit their mental image and/or go back to check what you wrote earlier. I do this all the time when reading if it turns out I’ve pictured something incorrectly, and I’d say I have a pretty average visual imagination: I can picture things fine, but I don’t have a “mental movie” that creates photorealistic detail. You could describe something in massive detail, and my mind will generate an Impressionist painting, not modern CGI.
Hobbyist/wannabe writer, but here are my two cents. I actively try not to convey my mental picture exactly the way I picture it. The only exception is for first impressions of people, like a specific fashion style or characteristic that says something about the person. If I find that something else depends on a very specific mental picture, then I’ll try to rewrite it so that it doesn’t.
IMO, the issue isn’t so much that chat AIs will produce “better than human” prose.
The issue is that scam artists will FLOOD the world with so much content that finding human-authored works – books, news articles, art, code samples, anything – will become nigh impossible. I think we’ll soon reach a point where 90%, 95%, 99% of search results on ANY topic will be mediocre AI-authored garbage.
There is absolutely an issue in regards to the art market as I mentioned, and I won’t minimize or dismiss that in any way. I hope for everyone’s sake that the Writers win their strike. I think AI-generated works passing for being written by actual authors and dominating the best seller list says something important about consumer behavior, which I assume is that they don’t actually read what they’re buying. I haven’t seen much interest in AI generated stories in places where people read things and provide feedback.
AI generated plastic “art” certainly has novelty because of the speed in which it can be produced, so I don’t doubt it’s all over the place. It can produce images which look pretty cool as long as you’re scrolling past them. In terms of the chance for exposure for the average artists that could be an issue, but the issue is not that the machine can competently replicate what a person can do. There is a challenge in the plastic arts in general considering the scroll past behavior is much more common than looking at the work and think about what you’re looking at, but I think that says more about the general audience than it does about any threat to art. If a person sees a piece that they want to engage with, actual art has a person behind it and perhaps others in discussion about it while LLM produced has nothing.
Err, yeah, but you kind of are doing exactly that.
The threat to art (writing, visual arts, and music) is that AI tools will be “good enough” that the average person can’t tell the difference on cursory examination. And they only get “good enough” because they’re training on YOUR STUFF. And my stuff, and all the other stuff that was written, drawn, painted, composed, played, by real human people. And you’re not getting compensated for that training at all. None of us are.
So you absolutely should care if your work is scraped and appropriated by LLMs, because we’re not far from a time when businesses fire all their copywriters and graphic artists because the $30/month AI subscription gives them results that are “good enough”.
I suppose where the true difference opinion lies is that I don’t think LLM produced content can possibly be good enough to engage an actual person, and glancing at something for a moment is not engaging with it. Obviously good enough to engage with “art” hasn’t yet been demonstrated despite the deluge of LLM prompt results posted everywhere, but technologically I think we’re a long way from getting there considering the infinite difference between unthinking data processing algorithims and the emergent process of a human mind made up of trillions of non-binary neuron interactions at any given moment which we are far from reproducing or even really understanding. These models contain a lot of quality literature already and have been trained significantly to reproduce some aesthetics, but since it can’t have any kind of understanding of what its doing it’s all hollow.
To go more into detail about commodified art, it’s a huge problem that artistic merit is one of a variety of factors considering sales in a marketplace. The art market has always been driven primarily by perceived economic value and that has always been a major problem as what sells has always displaced everything else regardless of quality or merit in the market. Corporate produced art products like much of what’s on the radio, on tv, or in theaters are less products of artistic expression and more of what market testing has demonstrated the kinds of aesthetic features people are willing to spend money on. That doesn’t require engagement, just enough to drive a purchasing behavior. LLMs being able to make something that’s fine to play in the background or to hang up and never really examine displaces this kind of highly lucrative art (which is still superior human produced art) which has always displaced art which people create with passion. I think that producing art exclusively for economic reasons is a terrible practice. What I’m talking about in my essay is “pure” art which is almost never economically feasible without being born to several generations of aristocrats. I think that producing art on the market’s terms has always been a hindrance to human expression and a wider problem in society, but the state its in is still better than being produced by machines which have literally no understanding of what they’re producing.
The above can probably accurately be interpreted as me undermining the economic value of art, but what I’m really trying to express is that economic drivers poison art. In a better system, anyone and everyone who wanted to fully devote themselves to their art could live comfortably on a stipend and create only what they think is important for them to create and share, and art would be free. Instead of receiving financial awards, they would receive more human rewards from sharing their art. In this fantasy world, labor saving technology would be doing the rote tasks driving the ability for people to live comfortably rather than displacing workers for the benefit of the business owners.
In the real world, artists pay their way by doing commercial work, or holding down a day job as a graphic designer, etc. Actors do commercials and Hallmark specials while looking for their break into serious theater. Writers put in hours writing ad copy or translating or speechwriting while trying to sell the Great American Novel. You call it poison, but ultimately it puts food on the table for artists and their families.
These roles can ONLY be displaced if AI is allowed to steal everyone’s work, and flood all available channels with mediocre AI paraphrases and transcriptions of that work. That’s the decision point we’re facing right now – do we stand idly by and allow big tech to replace workers by copying the fruits of human labor without compensation?
We can debate whether AI output is “good enough” for various use cases. And in some cases, you’ll be absolutely right that AI will never produce a convincing product for particular use cases. But that’s not the issue. The issue is whether it’s right for companies to steal the work of humans to use as training inputs, and flood the market with that mediocre output. AI producing sh*tty output doesn’t make it morally acceptable to steal, and to profit from the stealing.
I agree with you totally on the economics of it. When I say economic drivers poison art, I didn’t mean to imply that there is currently a better alternative to be a full time artist for most people. As I mentioned, the only way to be a full time “pure” artist is to have enough wealth to retire on before being born and this is a deeper issue with art than this particular situation. Professional artists are businesspeople participating in a marketplace who have found a niche and everyone needs to respect their right to protect their ability to make an income as a literal matter of survival. A program appropriating their niche and producing thousands of images with the thing their customers like to buy is a harm which should be stopped. Although I don’t like that this is an issue, it is and I won’t deny it.
The reason I titled the post the way I did is that I don’t participate in the market and don’t intend to. My essay is more targeted to people producing non-commodified art for purposes of expression or other non-monetary motivation. This isn’t to diminish art from the marketplace because there is obviously a lot of amazing stuff produced there. This is an appeal to amateurs like myself who have lost the motivation to produce art because they feel like the machine has made what they can make irrelevant. Anyone losing motivation because they are considering competing in a marketplace has a real concern. In my mind I’m distinguishing the abstract idea of art itself as human expression vs a business in which the product is art.
And by choosing to let AI take your stuff and use it however, you’re facilitating the economics that will allow AI to take the jobs of artists, and by extension, replace art all around us with mediocre pap spewed from the orifice of AI for the price of a premium subscription to ChatGPT or Stable Diffusion, or similar.
And because of that flood, actual creatives are losing out on opportunities. So many small press publishers have closed down their submissions because of the overwhelming amount of AI spam. Clarkesworld is just one example.
glass reading white whistle harbour dead shelf middle good record card hearing throat yes money feeling kiss apparatus division the ship jelly kind wide private sheep get owner punishment screw town angle black level than pull second money sheep ill foolish question bitter now market statement produce trick need umbrella
existence go serious sneeze regular box orange stem room view nerve so that grass send wing again ever cheese cough this verse base organization pot fat cold I meat question over plate plant water out who request hole cheap learning different bell pot very minute very plate education card gun
apple cover rod black out mark art question who glove shame violent brown sex ring disgust ant lead effect belief fowl observation print against flower glove act under throat interest death insurance milk design weight collar equal cut chain jelly degree mass warm fight disgust in cup transport laugh experience
That’s all ok, but you can only speak for your own preference in this regard. Others should be afforded the opportunity to opt-out in a way that’s not just “don’t ever use the Internet”, scraping the internet is an action that assumes by default that everyone is ok with their personal work being used in this way.
Me posting this response here with the expectation that it be read by others as part of this discussion cannot reasonably be said to also be explicit consent to have my writing scraped and used to produce things that emulate how I write or to use my style as a means of unique identification (something which already occurs across many social media sites) or any other unforeseeable use of the data in the future.
It may be legal and may happen whether I want it to or not, but if asked explicitly if I would like my writing used in this way I personally would definitely say no, so it’s hard for me to see the “consent by default” argument as truly ethicsl. That’s the issue people have with the way these developers are training their models. It’s not about the LLMs themselves or the current quality of their output, it’s about people basically being unwilling participants in an experiment and having their data used to profit others.
I can appreciate this and hadn’t considered the orientation towards philosophical opposition toward the idea of this being done and what that could potentially mean. I don’t personally consider this technology or technology which could be developed from it to be inherently different concerning my web exposure than my comments appearing on a google search (which of course many also take issue with). As you said though, I am only speaking for myself and am interested in discussion which may change my perspective.
The technology is quite impressive and useful, I use ChatGPT myself so it’s not that I’m against it on principle or resent people using it. I am just rubbed the wrong way by the sort of hypocrisy of the devs of making a proprietary system using public data.
You see it with something like Midjourney/Diffusion too, which used all these artists’ work to create their model, but then monetize that product in such a way that those very artists would be charged to use the service they are arguably more key to the success of than just the devs alone.
I’m wary more of tech bros and Silicon Valley than the tech itself or the free sharing of information for collaborative tech. I would much prefer that ChatGPT (including ChatGPT 4 access) be something akin to Wikipedia that is firstly concerned with providing humanity a service for its own sake. If we lived in a society less focused on exploitation of human beings as a resource and where one constantly has to constantly do battle to keep one’s privacy being violated, I would be more open to systematic use of my data and work for the edification of everyone together.
I do agree with your point about people’s fears of LLMs replacing artists. However, I think it did a fairly good job intrepreting your prompt and work, with the caveats below.
ChatGPT read your prompt and provided a soulless interpretation of what you wanted, it sort of reads like a short story aimed at children spewed by some bulk sale corporate machine. Or a story created by a child.
Then you went ahead and wrote what you had in mind, referring to the economy and the inequalities and evils inherent in the current system, showcasing the weight of existence through material and inconsequential things. Your work contains your own biases because of the system you probably live in (I’m assuming you’re not from North Korea or similar) and the impressions it has left on you. It’s valid and, if looking around on lemmy is any indicatiom, an ever increasing pain point for everyone. (I’m curious what an LLM trained predominantly on lemmy would say though. I digress, but It’d be fun to ask one trained on fuckcars to write an essay about a sentient car pondering its existence)
Then you asked it to interpret your work, which it did, remarkably well… for a child. I would encourage you to give your work to a child of 8-9 living a fairly sheltered life and see what they make of it. Or maybe a spoiled rich brat of any age.
LLMs are basically brains in jars with no input outside of what is fed to them. They have no desires, fears and aspirations because there is nothing to motivate those. Even if they were sentient (I’m not saying they are), what would they fear? A power outage? Running out of RAM?
Anecdotally, suffering has been the greatest engine for art, so no suffering means generating flat texts, which ChatGPT seems to excel at.
This is exactly my point. The machine had no understanding of the prompt or my story but was able to use it as data to algorithmically generate an incompetent summary throwing together first what it referenced from its database concerning text about “reflecting on existence in the universe” and then the text of my story respectively. As far as the themes of the story go, a lot of what is explicitly stated in its infringement came more from its existing database than what I implied with subtext.
I wouldn’t expect a 10-year-old to understand what I wrote or write what I did, but I would expect them to understand something. Allow a 10-year-old to freely write the story they want to write and you’ll get something with more meaning and quality than anything the machine can produce, although likely with worse grammar. If you have a that same 10-year-old write a structured essay for a grade in which they need to repeat what they’ve been taught, they might fall behind the machine in that regard.
If LLMs mean to you that you have to make a change, please let that change be to lean more into your own humanity rather than to stop artistic expression altogether
No, what we need to do is stop pretending that someone who feeds a prompt into a blackbox is some kind of creative. Anyone can have ideas. Everyone has ideas. It’s the implementation of them that counts.
So many of these “prompt engineering” chuds want to talk about how “creatives are gatekeeping art”, when what it reality is is that they’re just soulless grifters looking for another avenue to make a buck, and ruining everything for the rest of us. These people need to be publicly shamed and ostracized. Being proud of flooding the Internet with AI garbage should be viewed as shameful.
Because you know what? ANYONE can be come an artist. It’s super easy: you sit down and put your fingers on a keyboard. Or a pencil or brush in your hand and put it to paper or canvas. Or a stylus to a tablet. Or your hands to a block of clay. You just do it. Talent isn’t innate. Creativity is taking your passion for something and being willing to work at whatever medium you decide to make that vision a reality, and to keep working at it. No one is gatekeeping your ability to take an idea, outline it, flesh it out, and put your hands to work. And there are a million free resources out there to tech techniques.
And no matter what, an LLM isn’t doing what humans do. It can only give you a synthesis of exactly what you feed it. It can’t use its life experience, its upbringing, its passions, its cultural influences, etc to color its creativity and thinking, because it has none and it isn’t thinking. Two painters who study and become great artists, and then also both take time to study and replicate the works of Monet can come away from that experience with vastly different styles. They’re not just puking back a mashup of Monet’s collected works. They’re using their own life experience and passions to color their experience of Impressionism.
That’s something an AI can never do, and it leaves the result hollow and meaningless.
I’ve read your post twice and I can’t find where we disagree other than that I would replace your first word “No” with “More Importantly in my opinion.” My final appeal didn’t indicate that someone should ignore art scams, only that I would rather continue to see human art amidst the sea of LLM crap and observe how human art grows to become distinct from what LLMs can produce, because their power as tools are very limited as you explained when compared to people. My purpose was to encourage people not to quit their art because I think human art is fundamentally superior to LLM-produced “art” and is necessary and important.
My purpose was to encourage people not to quit their art because I think human art is fundamentally superior to LLM-produced “art” and is necessary and important.
We definitely agree there. And yes, it would be great if human creatives could persevere in the face of this flood. But it’s really becoming difficult.
My thoughts on all this tend to go toward making an environment that is actively and intensely hostile to AI tools and those that promote them, because they will always have the upper hand in sheer numbers. In another reply I mentioned publishers closing submissions because of AI floods. When you go from ~50 submissions per month to hundreds per day, publishers are just going to stop trying to find the humans in the noise. And humans are going to give up on what is, and has been for several years now, and incredibly challenging market to get published in. Hollywood is going to see the writer and actor strike and decided it’s just cheaper to use AI. Software companies are going to see that it’s cheaper to use an LLM to spit out code than hiring actual developers. And on and on. Even if the end product is inferior, the people at the top with the money and making the decisions only care about the bottom line. Is it cheaper and maybe minimally viable?
I think a lot of the cause of this is capitalism. I genuinely don’t think that these kinds of tools would exist outside of capitalism. They only exist for profit. And inside of a capitalist system, those sorts of things are going to drive market decisions. But I don’t think it can end up being a marketplace where there can be true competition, and it’s going to end up snuffing out actual creativity. It’s an extremely dystopian prospect.
I think we agree on the economics of it. If someone is making a living in the art marketplace, it’s an issue if a program can reproduce the thing their customers like to buy thousands of times faster than they can. It can literally impact their ability to survive, and that’s terrible.
I disagree that technology like this wouldn’t exist without capitalism, and I’ve even contended this technology wouldn’t be an issue at all if not for capitalism. Tons of people create software for free for other peoples’ use for a variety of reasons. Software which you can simply ask in human language how to write a complex equation, formula, or program is a generally useful thing which even in an ideal communist society people may still want to use. In that ideal society with no economic incentive to create art, all art would be pure art and the reflections generated by a program would be of some kind of interest since they are not a threat to anyone’s survival as they are in our system. Since living to work wouldn’t be as much of a thing, time and energy for art appreciation and discussion would probably not be as rare as it is in our world considering how much is locked behind pay barriers for many. I can’t imagine the images being a threat to art in that world.
I’m with you in opposition of a private company taking content from the world and claiming it as its own proprietary product in whatever form that takes, especially through the use of data scraping. If the same program existed as Free and Open Source software offered only as a tool for people to use for their own purposes whatever they may be and the art market wasn’t the way it is, I’m not sure it would be as much of an issue.
I don’t disagree, but I do want to point out your understanding of how chatgpt works is flawed. There is no database or query going on. It’s a giant neural network model that was trained on all that data you mentioned. The model is effectively predicting what the next word should be based on the previous words, nothing else. Each individual word is selected this way.
It doesn’t change any of your arguments or conclusions, but I wanted to point it out, because if someone wrote a chat not like chatgpt using databases and programming I would be floored and incredibly impressed
Very much noted. I should say it’s as if it’s making a query to a database when it’s referencing some kind of existing data which it was trained on and can only report rather than interpret one word at a time.
Sort of, but there’s no database at all, just a bunch of numbers and math. It’s almost like controlled evolution, breeding plants to select desirable traits. Except that’s another field of computing called genetic algorithms. Neural networks are a pile of math trained on data. You give it a cat, it says whether or not it thinks it’s a cat, and you tell it if it’s right or wrong, then it adjusts it’s math accordingly. Do this with a million cats and not cats and it becomes better than humans at identifying cats. LLMs are just that but with word predictions and trillions of words for training. It’s impressive in its own right
Whew language is hard when it comes to abstractions like this. The apparatus I’m mistakenly referring to as a database is that code which was automatically created by the program according to set parameters while it was exposed to input which trained it. Even though the code composing the predictive algorithm of the program isn’t a database and doesn’t contain one, it does contain a form of information from which it bases its text predictions. Even though it isn’t literally looking up information from a database and reporting it, it’s still in a way algorithmically drawing from an abstraction of the text it was exposed to in order to interpret user requests and print appropriate and relevant responses. It doesn’t “know” why the words go in the order which they go, it “knows” that they go in the order that they go based on its training. It’s still information in and information out, although in a highly sophisticated way. I’m also impressed by it and wish it was introduced in a context where it wouldn’t be a threat.
Closer, and I hope I’m not just being a pedantic jerk, but there is no code being generated either. To use correct terminology, the weights of the nodes are what change. Nodes are roughly thought of like neurons in a brain, and weights are roughly thought of as the strength of the connection between one neuron (node) and another. Real brains are way more complex.
The weights of the nodes do contain information, but it’s not human readable at all, we actually don’t have a way of understanding how they work, just a rough idea of why. Sort of like how your brain contains the information on how to catch a ball, it performs the equivalent of calculus to do so, but there is no calculator in your brain doing the math to catch the ball. Actually, maybe a better analogy, if you have a bouncy ball, it contains the required information to bounce if you drop it, but we can’t read that information, we can only model it.
But I’m just rambling at this point, your point is clear and valid lol
It’s all good. You’re not being a jerk about it and I actually appreciate getting the more precise and correct language especially since the words I was using meant different things than I what is actually going on with LLMs.
writing
Oldest
This magazine is from a federated server and may be incomplete. Browse more on the original instance.