4 ms·

in terms of sheer spam, they might be killing the sense of awe in someone who found them awe-inspiring to begin with. i can't speak for anyone else but i don't

in terms of sheer spam, they might be killing the sense of awe in someone who found them awe-inspiring to begin with. i can't speak for anyone else but i don't exactly find myself seeking out stable diffusion output because a big colourful high-contrast image of a t. rex bursting out the center of mars surrounded by rainbow mandala constellations or whatever doesn't really do anything for me. it's just blockbuster clip art. we had already been living in a broader culture that desensitises people to the impression of substance prior to dall-e or whatever; netflix originals and nolan flicks are visibly high-production-value but, generously, one out of every ten actually says anything. even entirely hand-produced "blockbuster clip art" like that one pic of the cat with big round sunglasses with space reflecting in it that's been a fixture of 4chan wallpaper threads since time immemorial has long since served as an emblem of the easily-amused (see also any visual artifact that is a mainstay of stoners). and all it is is an original photograph by a certain андрей прохоров from a series known as "a girl and her cat", subject to a deviantart user's "10 minutes in photoshop for practice" for the reflections. the real risk to awe is highlighted further down in the article and touched upon far too briefly relative to the firehose thing, which is the stain upon awe of suspicion. is this image from a camera or a 3d package or a generative neural network? balancing a healthy fascination with the world with a healthy suspicion of your senses is a discussion that risks veering into dangerously political, religious, not-very-hacker-news-guidelines territory, but i'll play it safe and just say an apocalyptic tone is not conducive to determining what the healthy balance is.

soulofmischief 3y ago
In your denouncement of the ostensibly trite content of "a t. rex bursting out the center of mars surrounded by rainbow mandala constellations or whatever", you fail to look past the content itself and find value in the technique. From an artist's perspective, these models are nothing short of magical in how they manage to pack in knowledge across so many styles and contexts. It's personally mind-blowing how effective diffusion is in general (but retrospectively, it's a beautifully simple concept), but it's not just the isolated image of a t-rex on mars that it interesting. It's the speed and cost at which the image can be generated, and the ways in which the data can be modified or recontextualized in a feedback loop. Or how multimodal systems like GPTV or LLaVA allow a new form of communication which blurs the line between images and language. I could go on, but the essence is that you are doing yourself a severe disservice by generalizing these capabilities as "blockbuster clip art".
- 0x69420 3y ago
  i met the subject matter on the article's terms and made it quite explicit that an image can be impressive but unsubstantial entirely orthogonal to whether or not it is the result of diffusion. obviously it's nuts that you can cook up the right markov chain to repeatedly denoise your way to something recognisable; i doubt anyone's here to take that away from you.
- itronitron 3y ago
  No one's arguing that the people using generative AI are enthralled with the methods, and I expect that people that make spin art are equally enthralled with the process and also amazed at the images they create, but I wouldn't call either art.
  - Teever 3y ago
    What if we didn't get into a semantic argument about the word 'art'?
    - itronitron 3y ago
      Just call it 'math generated imagery' because that is what it is.
      - pixel8account 3y ago
        What about "Gimp generated imagery" for computer graphic, and "brush generated imagery" for oil paintings?
  - soulofmischief 3y ago
    The coolest thing about art is that you don't get to decide what it is.
    - __loam 3y ago
      You're still a hack if all you do is prompt stable diffusion.
civilitty 3y ago
> i can't speak for anyone else but i don't exactly find myself seeking out stable diffusion output because a big colourful high-contrast image of a t. rex bursting out the center of mars surrounded by rainbow mandala constellations or whatever doesn't really do anything for me. Oh yeah? What about... "a mesmerizingly vibrant and colossal T-Rex with an iridescent, shimmering skin that reflects all the colors of the spectrum. It stands in a fantastical, alien landscape filled with towering, luminescent mushroom-like structures and a flowing river of brilliant, neon light. In the background, an otherworldly nebula explodes with the most intense and vivid colors across the cosmic star-filled sky"!? [1] I have nothing else to add to this serious topic, I just like drawing dinosaurs like I'm eight years old again [2]. I rented a GPU for this weekend and I've got a script using GPT4 to generate random topic-specific prompts for SDXL to test out all the parameters, schedulers, etc (5k images and counting so far). Maybe it's just the fact that I've avoided the reddit firehose but I don't know how anyone could lose their awe so fast. It's just fun and there doesn't have to be any more to it. [1] https://dalle.party/?party=cA8EVQF5 https://dalle.party/?party=cA8EVQF5 [2] https://dalle.party/?party=pXWEenFO https://dalle.party/?party=pXWEenFO
- 0x69420 3y ago
  > I have nothing else to add to this serious topic despite your best efforts, you've done so in the following: > Maybe it's just the fact that I've avoided the reddit firehose > It's just fun these really are important variables! you have the process, you have two models playing a game of telephone, you have to hold the fact of the models' existence in your head to appreciate that page. that's so much more than just one of those images out of context popping up on a social media feed.
injidup 3y ago
A friend of mine built a snowman today. It was beautiful. No photo real or generated can do it justice. The act was in the building and being there. That is the key takeaway from all this. Choose real life. Go out and build a snowman. Interact with the world and with people and materials. It doesn't matter if a master photographer or an AI generates a picture of a snowman. In the end if it motivates you to go out and build one yourself then it is good. Otherwise it is just porn.
- epivosism 3y ago
  My ggggg*1000 grandfather, a eukaryote, really knew how to live. He didn't have hands - he had to DREAM and use his MIND to make snowmen. Kids today with their "hands" are desensitized to the needed mental work of imagining how a snowman would look. I'm not saying this argument automatically wins. But I think the two sides need to define terms. What's "real" and "authentic" for one generation was a huge stretch for another, and will be rote/boring for another. Is there a coherent dividing point? I personally don't think so; I would like to hear an attempted definition, though. Real = Struggling to manage your horse to plow the fields so you and your family doesn't starve? Being a master horseman and ploughman? Breeding sturdier horses and selling them so you're rich, and everyone can easily plow? Replacing horses with engines+plows, with heated cabins, which activates way more land, to grow more food for the world? Running a retro amish-style farm on the weekend where you go back to using real horses to plow again? ^^^^ Where does "real/genuine" come in and then leave in that sequence? And who are you to say where other people should spend their time within it?
vorticalbox 3y ago
> because a big colourful high-contrast image of a t. rex bursting out the center of mars surrounded by rainbow mandala constellations https://i.postimg.cc/KcHPQSmf/image.png https://i.postimg.cc/KcHPQSmf/image.png
- telesilla 3y ago
  >it's just blockbuster clip art Well said, op
davesque 3y ago
Agree. Another example of desensitization is the popularity (now diminishing) of Marvel movies. Relative to their visual presentation, they are empty of content. Gauging the depth of meaning in a Marvel movie takes almost no effort. I suppose their recent drop in popularity might indicate that the broader culture is finally catching on to this.
- skydhash 3y ago
  The Loki series and the Spider-Verse animations were a bit of fresh air in the superhero space. I watched The Marvels recently and it’’s just visual, nothing else.
- Animats 3y ago
  That's true of the entire genre. I've heard of acting classes where people tried doing a dramatic reading from Star Wars. Without the special effects and music, it falls flat.
- pixel8account 3y ago
  As someone who watched all the "original" marvel movies, up to a infinity war I think, I don't think anyone expected any depth from them . I wanted to watch a classical sorry full of non-subverted tropes about a hero fighting an adversary, going through a classical hero journey, and eventually prevailing (possibly sacrificing something in a process). This is a tale as old as time and I don't think there's anything wrong with it. Of course marvel movies are devoid of any greater meaning and don't really prompt the viewer to think about anything deeper, but for me it's a feature, not a bug - when I want to think I open a technical book or start a side project.
chongli 3y ago
Yeah, if you want to be awed, go outside. Visit a forest after the rain. Or climb a big hill and look around. Or use a light pollution map to find a nice dark sky to go out and see the stars (admittedly much harder than it used to be). Nature is awesome. Most photographs don’t do it justice. You have to get out there and see it for the full sensory experience.
- saiya-jin 3y ago
  Nature is full sensory experience and much more (physical effort, rewards). Also, we evolved to have it as a natural baseline, and folks in cities just vaguely lost it over few generations (or one), its still there if you actually listen to your body and mind. The only people that I ever met that dislike nature and mountains are severely depressed people (and not all of those of course, and even those often flat out reject objectively helping things), literally everybody else heals a bit or recharges there.
ilaksh 3y ago
If it's meaning you want from art, then DALLE-3 or SDXL plus extensions or abilities like inpainting might actually be the most effective way to achieve that. After all, there is quite a bit of meaning in the text used to generate them. That seems like a head start if you are trying to make a statement. Not to mention the speed that the image or image elements emerge and can be iterated or integrated. For example, take Judy Chicago's The Dinner Party (admittedly just from the top of Google search results about artistic statements). It is a series of plates that have sculptures depicting femininity in various ways at various heights. With something like a painting tool with SDXL or DALLE-3 integrated one could create an image with similar content by drawing the table and then generating/inserting or inpainting each of the individual plate contents per specification such as the size of the inpaint mask and prompt. She clearly explained the meaning of the work and it can fit in a few long sentences. The more I think about this, the more it seems that image generation should be very helpful for artists who want to "say" something.

soulofmischief 3y ago

In your denouncement of the ostensibly trite content of "a t. rex bursting out the center of mars surrounded by rainbow mandala constellations or whatever", you fail to look past the content itself and find value in the technique. From an artist's perspective, these models are nothing short of magical in how they manage to pack in knowledge across so many styles and contexts. It's personally mind-blowing how effective diffusion is in general (but retrospectively, it's a beautifully simple concept), but it's not just the isolated image of a t-rex on mars that it interesting. It's the speed and cost at which the image can be generated, and the ways in which the data can be modified or recontextualized in a feedback loop. Or how multimodal systems like GPTV or LLaVA allow a new form of communication which blurs the line between images and language. I could go on, but the essence is that you are doing yourself a severe disservice by generalizing these capabilities as "blockbuster clip art".

0x69420 3y ago
i met the subject matter on the article's terms and made it quite explicit that an image can be impressive but unsubstantial entirely orthogonal to whether or not it is the result of diffusion. obviously it's nuts that you can cook up the right markov chain to repeatedly denoise your way to something recognisable; i doubt anyone's here to take that away from you.
itronitron 3y ago
No one's arguing that the people using generative AI are enthralled with the methods, and I expect that people that make spin art are equally enthralled with the process and also amazed at the images they create, but I wouldn't call either art.
- Teever 3y ago
  What if we didn't get into a semantic argument about the word 'art'?
  - itronitron 3y ago
    Just call it 'math generated imagery' because that is what it is.
    - pixel8account 3y ago
      What about "Gimp generated imagery" for computer graphic, and "brush generated imagery" for oil paintings?
- soulofmischief 3y ago
  The coolest thing about art is that you don't get to decide what it is.
  - __loam 3y ago
    You're still a hack if all you do is prompt stable diffusion.

civilitty 3y ago

> i can't speak for anyone else but i don't exactly find myself seeking out stable diffusion output because a big colourful high-contrast image of a t. rex bursting out the center of mars surrounded by rainbow mandala constellations or whatever doesn't really do anything for me. Oh yeah? What about... "a mesmerizingly vibrant and colossal T-Rex with an iridescent, shimmering skin that reflects all the colors of the spectrum. It stands in a fantastical, alien landscape filled with towering, luminescent mushroom-like structures and a flowing river of brilliant, neon light. In the background, an otherworldly nebula explodes with the most intense and vivid colors across the cosmic star-filled sky"!? [1] I have nothing else to add to this serious topic, I just like drawing dinosaurs like I'm eight years old again [2]. I rented a GPU for this weekend and I've got a script using GPT4 to generate random topic-specific prompts for SDXL to test out all the parameters, schedulers, etc (5k images and counting so far). Maybe it's just the fact that I've avoided the reddit firehose but I don't know how anyone could lose their awe so fast. It's just fun and there doesn't have to be any more to it. [1] https://dalle.party/?party=cA8EVQF5 https://dalle.party/?party=cA8EVQF5 [2] https://dalle.party/?party=pXWEenFO https://dalle.party/?party=pXWEenFO

0x69420 3y ago
> I have nothing else to add to this serious topic despite your best efforts, you've done so in the following: > Maybe it's just the fact that I've avoided the reddit firehose > It's just fun these really are important variables! you have the process, you have two models playing a game of telephone, you have to hold the fact of the models' existence in your head to appreciate that page. that's so much more than just one of those images out of context popping up on a social media feed.

injidup 3y ago

A friend of mine built a snowman today. It was beautiful. No photo real or generated can do it justice. The act was in the building and being there. That is the key takeaway from all this. Choose real life. Go out and build a snowman. Interact with the world and with people and materials. It doesn't matter if a master photographer or an AI generates a picture of a snowman. In the end if it motivates you to go out and build one yourself then it is good. Otherwise it is just porn.

epivosism 3y ago
My ggggg*1000 grandfather, a eukaryote, really knew how to live. He didn't have hands - he had to DREAM and use his MIND to make snowmen. Kids today with their "hands" are desensitized to the needed mental work of imagining how a snowman would look. I'm not saying this argument automatically wins. But I think the two sides need to define terms. What's "real" and "authentic" for one generation was a huge stretch for another, and will be rote/boring for another. Is there a coherent dividing point? I personally don't think so; I would like to hear an attempted definition, though. Real = Struggling to manage your horse to plow the fields so you and your family doesn't starve? Being a master horseman and ploughman? Breeding sturdier horses and selling them so you're rich, and everyone can easily plow? Replacing horses with engines+plows, with heated cabins, which activates way more land, to grow more food for the world? Running a retro amish-style farm on the weekend where you go back to using real horses to plow again? ^^^^ Where does "real/genuine" come in and then leave in that sequence? And who are you to say where other people should spend their time within it?

vorticalbox 3y ago

> because a big colourful high-contrast image of a t. rex bursting out the center of mars surrounded by rainbow mandala constellations https://i.postimg.cc/KcHPQSmf/image.png https://i.postimg.cc/KcHPQSmf/image.png

telesilla 3y ago
>it's just blockbuster clip art Well said, op

davesque 3y ago

Agree. Another example of desensitization is the popularity (now diminishing) of Marvel movies. Relative to their visual presentation, they are empty of content. Gauging the depth of meaning in a Marvel movie takes almost no effort. I suppose their recent drop in popularity might indicate that the broader culture is finally catching on to this.

skydhash 3y ago
The Loki series and the Spider-Verse animations were a bit of fresh air in the superhero space. I watched The Marvels recently and it’’s just visual, nothing else.
Animats 3y ago
That's true of the entire genre. I've heard of acting classes where people tried doing a dramatic reading from Star Wars. Without the special effects and music, it falls flat.
pixel8account 3y ago
As someone who watched all the "original" marvel movies, up to a infinity war I think, I don't think anyone expected any depth from them . I wanted to watch a classical sorry full of non-subverted tropes about a hero fighting an adversary, going through a classical hero journey, and eventually prevailing (possibly sacrificing something in a process). This is a tale as old as time and I don't think there's anything wrong with it. Of course marvel movies are devoid of any greater meaning and don't really prompt the viewer to think about anything deeper, but for me it's a feature, not a bug - when I want to think I open a technical book or start a side project.

chongli 3y ago

Yeah, if you want to be awed, go outside. Visit a forest after the rain. Or climb a big hill and look around. Or use a light pollution map to find a nice dark sky to go out and see the stars (admittedly much harder than it used to be). Nature is awesome. Most photographs don’t do it justice. You have to get out there and see it for the full sensory experience.

saiya-jin 3y ago
Nature is full sensory experience and much more (physical effort, rewards). Also, we evolved to have it as a natural baseline, and folks in cities just vaguely lost it over few generations (or one), its still there if you actually listen to your body and mind. The only people that I ever met that dislike nature and mountains are severely depressed people (and not all of those of course, and even those often flat out reject objectively helping things), literally everybody else heals a bit or recharges there.

ilaksh 3y ago

If it's meaning you want from art, then DALLE-3 or SDXL plus extensions or abilities like inpainting might actually be the most effective way to achieve that. After all, there is quite a bit of meaning in the text used to generate them. That seems like a head start if you are trying to make a statement. Not to mention the speed that the image or image elements emerge and can be iterated or integrated. For example, take Judy Chicago's The Dinner Party (admittedly just from the top of Google search results about artistic statements). It is a series of plates that have sculptures depicting femininity in various ways at various heights. With something like a painting tool with SDXL or DALLE-3 integrated one could create an image with similar content by drawing the table and then generating/inserting or inpainting each of the individual plate contents per specification such as the size of the inpaint mask and prompt. She clearly explained the meaning of the work and it can fit in a few long sentences. The more I think about this, the more it seems that image generation should be very helpful for artists who want to "say" something.