• 1 Post
  • 257 Comments
Joined 2 years ago
cake
Cake day: June 12th, 2023

help-circle





  • I’m afraid that we seen to disagree on who an artist is and what is a valid moral trade off.

    Is it really the democratization of art? Or the commodification of art?

    Art has, with the exception of extraordinary circumstances, always been democratic. You could at any point pick up a pencil and draw.

    Ai has funneled that skill, critically through theft, into a commodified product, the ai model. Through with they can make huge profits.

    The machine does the art. And, even when run on your local machine the model was almost certainly trained on expensive machines through means you could not personally replicate.

    I find it alarming that people are so willing to celebrate this. It’s like throwing a party that you can buy bottled Nestle water at the grocery store which was taken by immoral means. It’s nice for you, but ultimately just further consolation of power away from individuals.


  • Sorry, I might have went a bit ham on you there, it was late at night. I think I might have been rude

    1. Theft does not depend on a legal definition.

    Intellectual property theft used to be legal, but protections were eventually put in place to protect the industry of art. (I’m not a staunch defender if the laws as they are, and I belive it actually, in many cases, stifles creativity.)

    I bring up the law not recognizing machine generated art only to dismiss the idea that the legal system agrees wholeheartedly with the stance that AI art is defensibly sold on the free market.

    1. There is no evidence to suggest AI think like a human / It hardly matters that AI can be creative.

    A) To suggest a machine neutral network “thinks like a human” is like suggesting a humanoid robot “runs like a human.” It’s true in an incredibly broad sense, but carries so little meaning with it.

    Yes, ai models use advanced, statistical multiplexing of parameters, which can metaphorically be compared to neurons, but only metaphorically. It’s just vaguely similar. Inspired by, perhaps.

    B) It hardly matters if AI can create art. It hardly even matters if they did it in exactly the way humans do.

    Because the operator doesn’t have the moral or ethical right to sell it in either case.

    If the AI is just a stocastic parrot, then it is a machine of theft leveraged by the operator to steal intellectual labor.

    If the AI is creative in the same way as a person, then it is a slave.

    I’m not actually against AI art, but I am against selling it, and I respect artists for trying to protect their industry. It’s sad to see an entire industry of workers get replaced by machines, and doubly sad to see that those machines are made possible by the theft of their work. It’s like if the automatic loom had been assembled out of centuries of collected fabrics. Each worker non consensually, unknowingly, contributing to the near total destruction of their livelihood. There is hardly a comparison which captures the perversion of it.


  • Counterpoints:

    Artists also draw distinctions between inspiration and ripping off.

    The legality of an act has no bearing on its ethics or morality.

    The law does not protect machine generated art.

    Machine learning models almost universally utilize training data which was illegally scraped off the Internet (See meta’s recent book piracy incident).

    Uncritically conflating machine generated art with actual human inspiration, while career artist generally lambast the idea, is not exactly a reasonable stance to state so matter if factly.

    It’s also a tacit admission that the machine is doing the inspiration, not the operator. The machine which is only made possible by the massive theft of intellectual property.

    The operator contributes no inspiration. They only provide their whims and fancy with which the machine creates art through mechanisms you almost assuredly don’t understand. The operator is no more an artist than a commissioner of a painting. Except their hired artist is a bastard intelligence made by theft.

    And here they are, selling it for thousands.


  • Flirting is a pretty nebulous term.

    My personal definition of flirting is any positive expression or behavior which:

    1. Is an exception to your typical behavior or affect.
    2. Targeted at a specific person, typically someone new.
    3. Is heightened or marked by increased volume, nervousness, etc.

    examples:

    A typically reserved guy pulling you aside and animatedly asking about your interest.

    A woman who normally doesn’t touch you is repeatedly teasing you about your shirt, pulling on the fabric.

    A good friend begins to repeatedly and unexpectedly invite you over for one on one movie nights despite obvious inconveniences.

    That’s just my opinion, though. I believe most people are looking for these heightened expressions of flirting to confirm interest.



  • Perhaps they assumed that you felt obligated to speak to them, and they felt the need to absolve you of that obligation. That would be the literal meaning of the phrase.

    A different interpretation could be that they didn’t want to continue the conversation. In this scenario, they would prefer to be silent, and by suggesting that, “You don’t have to do this,” they have given you a polite, symbolic agency in ending the conversation, implying that both of you would rather be silent. In this way, instead of directly rejecting you, the decision to be silent would be shared. It’s a way of allowing you to save face.

    Note that just because they might want to be silent, doesn’t mean that they dislike you. There are many reasons why people might know to speak at any given moment, and some people are just shy.

    It’s hard to say which interpretation is correct without hearing the tone of voice, and understanding the general situation. There could be other factors or other interpretations as well.


  • Yes, sorry, where I live it’s pretty normal for cars to be diesel powered. What I meant by my comparison was that a train, when measured uncritically, uses more energy to run than a car due to it’s size and behavior, but that when compared fairly, the train has obvious gains and tradeoffs.

    Deepseek as a 600b model is more efficient than the 400b llama model (a more fair size comparison), because it’s a mixed experts model with less active parameters, and when run in the R1 reasoning configuration, it is probably still more efficient than a dense model of comparable intelligence.




  • This article is comparing apples to oranges here. The deepseek R1 model is a mixture of experts, reasoning model with 600 billion parameters, and the meta model is a dense 70 billion parameter model without reasoning which preforms much worse.

    They should be comparing deepseek to reasoning models such as openai’s O1. They are comparable with results, but O1 cost significantly more to run. It’s impossible to know how much energy it uses because it’s a closed source model and openai doesn’t publish that information, but they charge a lot for it on their API.

    Tldr: It’s a bad faith comparison. Like comparing a train to a car and complaining about how much more diesel the train used on a 3 mile trip between stations.