top 50 comments

sorted by: hot top controversial new old
[–] 1 point 3 hours ago

the thing is, you give your thought to ai and ai generates what it thinks you mean through other peoples examples but you only have your idea of what you want. and you only have so many chances for ai to get it right

  • source
  • [–] 6 points 14 hours ago (1 child)

    “From a distance, Portrait of Edmond de Belamy, which has just sold at Christie’s in New York for $432,000 (£337,000), looks almost plausible. Up close, however, the paintwork becomes a grid of mechanical-looking dots, the man’s face a golden blur with black holes for eyes. Look into those eyes. They show no sign of feeling or life. Did a computer make this? The answer is yes. The first artwork generated by AI to be sold at Christie’s, its impressive price would seem to suggest that in future we will get computers to make art for us. Robot van Gogh will harmlessly cut its ear off and robot Picasso will be a genius, minus the misogyny. Is this the future AI art visionaries such as the French collective Obvious, which programmed this “painter” by getting it to compare its own work with 15,000 pre-20th century portraits, have in mind? Or are they just, God forbid, making a fast buck from gullible art collectors? Because believing the algorithm that knocked this up to be in any meaningful way an “artist” is like thinking your voice-interaction programme is out to get you. Dream on. Computers would need to replicate human consciousness before they could replicate the funny thing humans do called “art”. Art is a way in which human consciousness expresses itself, and is equally true of the earliest cave art, Rembrandt’s portraits and Duchamp’s urinal. And that is what is missing from Portrait of Edmond Bellamy. Art is a way humans communicate ideas, perceptions and feelings to each other. It has no existence outside the human passion to communicate. So in what meaningful sense can an AI replica of certain physical traits of old master paintings be called art? For a robot to really make art, it would need an autonomous mind that was emotional as well as rational. No AI developer has yet claimed to be anywhere near achieving that and if they ever do, their creation will probably have better things to do than paint portraits – like destroy humanity. Maybe afterwards robots will invent their own kind of art, but it won’t be some poor pastiche of human genius. It will be beyond anything we organics could imagine."

  • source
  • hideshow 1 child comment
  • [–] 3 points 13 hours ago

    But real art is not being sold for plebs with the promise of sticking it to "millionaire" artists that are all "full of themselves".

  • source
  • [–] 37 points 1 day ago (26 children)

    why don't we cut the middle man and just deal in prompts?

  • source
  • hideshow 26 child comments
  • [–] 2 points 4 hours ago

    This is something a good buddy of mine can't seem to grok. He loves generating ai stuff and showing it off and I keep telling him I love seeing his ideas, not what some corporate model decided to fill in based on his input.

    I am actually, truthfully more interested in seeing his idea, the raw prompt that he entered into his SaaS chatbot or even local LLM, than I am with sifting through the generated noise for the signal of the mind of the man I love.

  • source
  • parent
  • [–] 18 points 1 day ago (17 children)

    I'm sure this is meant as a joke. But, really, prompts are the only part of AI art that is actually creative. Especially if someone has been trying to fine tune what they want for a while, the prompt might actually have something interesting in it.

    Even though memes are shared as images, it's often the text that's the important and funny bit. The meme image is just context for the words.

    I'm not saying I'm actually interested in reading people's AI prompts. But, there might be an interesting idea in that prompt that the person could express in a way that's not boring and soulless like GenAI tends to be.

  • source
  • parent
  • hideshow 17 child comments
  • [–] 18 points 1 day ago (3 children)

    This came up back on Reddit in the "what's that book?" sub. Someone posted asking about a book; they had 6 lines of preamble about why they wanted to reread this, and 55 lines that were an excerpt from the book. This is normally a slam dunk--it's pretty trivial to search for literal text. But nobody could find it. Discussion with the user led to the revelation that they'd asked AI about a book and described a scene they remembered, and the AI had responded with a title and "excerpt" matching what the user described. But then the user still couldn't find the book, so they posted the "excerpt" to the subreddit. I.e., they were asking for a book containing an AI hallucination, rather than just describing the book to us the way they had to the AI. In fact the story they were searching for was "Harrison Bergeron," one of the most commonly asked-about things on the subreddit, but nobody recognized it until we understood that the excerpt was slop made up in response to a scene they misremembered as part of the story.

    It's a frustrating time to be a human on the internet.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 4 points 1 day ago (2 children)

    Man, that sucks. It's really annoying how sycophantic these companies have made their chatbots, which makes this worse.

    But, I think by the very nature of how chatbots work, you can't fix that problem. Chatbots simulate what text is likely to follow other text.

    If someone online has definitely read a book that really exists, and they describe a scene they remember from that book, the clear pattern online is that someone will say "Oh yeah, that's the book X by author Y". Some people might say "I've never heard of that book", but it's likely they might just not say anything. If we're looking at a kind of forum with upvotes and downvotes, the person who found the real book will be the most upvoted.

    So, that's the pattern the LLM trains on. It knows that when someone posts a question like that, the best answer is one where someone names the book and types out the passage in the book.

    It would be nice if the chatbot would instead say "I don't know that book" or "That book doesn't exist".

    But, it can't do that, because it doesn't "know" anything. Everything the chatbots put out is a hallucination. It's just that some hallucinations are more useful. It has been trained millions of times on patterns when the passage someone remembers is "It was the best of times, it was the blurst of times", so when it hallucinates about that sequence of words, its hallucination is bound to be more helpful. But, if what someone remembers doesn't exist, or is obscure enough that it's not in the training data, things break. Then it follows the general pattern used for every other book, and comes up with a "useful" answer that doesn't match any actual book.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 19 hours ago (1 child)

    It's definitely in the nature of the current-gen systems. It's possible to imagine other ways to train chatbots that wouldn't favor bullshitting in the same way, though. The way they currently are isn't a way they necessarily have to be.

    Anyway, it's very clearly a case where it would have been better to cut out the middleman and just share the prompt, which was my point in sharing it.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 4 hours ago

    It’s possible to imagine other ways to train chatbots that wouldn’t favor bullshitting in the same way, though

    How? The way they work is that they predict the next word (token). They don't know anything. So, when someone says "what's the book that features an X doing Y and has a Z that looks like a Q" the process is basically "if this were a forum post, what's the next word that's likely to appear, what's the one after that, what's the one after that".

    If the scene being described is something from Harry Potter that has been discussed endlessly on the web, it's likely to predict next words that end up being helpful. If the scenario being described has never been described before, there's no way for it to say "aha, this book doesn't exist!" Instead it just invents a plausible conversation between two people about a fictional book.

    Now, you could design a filter that sits in front of the LLM that makes things better. It could be designed to recognize quotes, and if it spots a quote, it can "search the web" (or really look through a big index) to see if those words in that sequence have actually been seen before. If not, it could reject the entire response as being a "hallucination".

    But, fundamentally, this isn't a problem you can fix through training, because of how LLMs work.

  • source
  • parent
  • load more comments (13 replies)
  • load more comments (3 replies)
    [–] 44 points 1 day ago* (last edited 1 day ago) (7 children)

    Something ive noticed about AI art is it enables people who see themselves as "idea guys" to pump out every thought they have, call it good enough, and display it proudly like they worked hard at it.

    When you make art you take an idea and that idea is shaped by external factors. Your skill level, your medium, how you collaborate if you are, all of these things influence what you make and that influence in turn changes your ideas. You may have an idea of what you want to paint, and during the process of painting come to appreciate the beauty of texture, and the next painting seeks to explore not just your idea but what you can do with the medium. Artists usually seek to push themselves to learn and explore and have their art change them as much as they seek to create the art.

    But an idea guy only has interest in the idea. He doesnt want to do the process, he has no desire to be changed by his work. He has an idea and he wants an end product, and AI allows him to bypass that pesky learning to draw/paint/create business. It removes limitations in medium and skill level and spits out something close enough to his idea that its serviceable.

    And I think that, to me, really signifies how much the idea guy values ideas over everything else. It takes all the work out of creation, and the lack of craftsmanship in the ideas is very telling in the final product.

  • source
  • hideshow 7 child comments
  • load more comments (6 replies)
    [–] 64 points 1 day ago (31 children)

    AI art also helps keep starving artists unfed

  • source
  • hideshow 31 child comments
  • load more comments (31 replies)
    [–] 130 points 1 day ago (22 children)

    is the accumulation of one's youtube feed truly the deepest gaze into the soul?

  • source
  • hideshow 22 child comments
  • load more comments (22 replies)
    [–] 75 points 1 day ago (2 children)

    I actually closed my eyes and took a moment to prepare myself after reading the first sentence because I knew I was about to read the most brain-dead take imaginable.

  • source
  • hideshow 2 child comments
  • load more comments (2 replies)
    [–] 47 points 1 day ago* (3 children)

    Engagement farming? I refuse to believe the top person is for real

    Also Youtube feed is the videos they haven't watched, so isn't in their mind (yet)

  • source
  • hideshow 3 child comments
  • load more comments (3 replies)
    load more comments
    view more: next ›