• godsammitdam@lemmy.zip
    link
    fedilink
    English
    arrow-up
    7
    ·
    3 hours ago

    Where’s all the people shouting how awesome capitalism is, private ownership, the communists will come steal everything from you.

    Here come capitalists to steal everything you own so they can make more money. Copyright, shmopyright. They’ve also got eminent domain on their side to build walls and data centers

  • GreenKnight23@lemmy.world
    link
    fedilink
    English
    arrow-up
    10
    ·
    5 hours ago

    gonna start my own darknet with a 56k bandwidth limit.

    shit was simpler 30 years ago because nothing advanced enough could use such a slow connection.

  • No1@aussie.zone
    link
    fedilink
    English
    arrow-up
    36
    ·
    7 hours ago

    A follow up question, please:

    Q: So, what is illegal for an AI firm to do?

      • myyass@lemmy.world
        link
        fedilink
        English
        arrow-up
        6
        ·
        6 hours ago

        this is worse than you can imagine. Started college again. Syllabus says, you can’t use AI. Okay, login to the school website; starts with microslops email login to straight to chat gpt. Okay, need to search for something on the school website, here’s the schools AI search. Okay, answer a topic but we will check and grade it with various AI platforms. Okay, write things in your own words and then I get a 0/100 for doing so. I answered it correctly, exponentially better than the outdated class. So yeah, they can get bent. The best part was that the other students that used LLMs to regurgitate my answer differently. Maybe one less person working in cybersecurity is overall better and I’ll just go back to fixing stuff and watch it all get cooked.

  • bss03@infosec.pub
    link
    fedilink
    English
    arrow-up
    2
    arrow-down
    1
    ·
    3 hours ago

    It’s mostly a nothing-burger right now:

    This new Trump administration brief is not a ruling, as the case is being tried in the U.S. District Court for the Southern District of New York, and the authors of the brief do not have jurisdiction.

    The U.S. copyright office somewhat contradicts the brief:

    The training process [also] implicates the right of reproduction.

    https://www.copyright.gov/ai/ Part 3, III, B.

    Tho, the same documents do mention that “fair use” defenses can apply, but as always they depend on a lot of factors specific to individual acts of possible infringement.

    • 4am@lemmy.zip
      link
      fedilink
      English
      arrow-up
      8
      arrow-down
      4
      ·
      edit-2
      6 hours ago

      I mean, based -

      But I don’t want the elimination of IP rights without a guaranteed income for artists.

      It’s not possible to eliminate IP rights under capitalism. Press the Socialism button!

      • NoneOfUrBusiness@fedia.io
        link
        fedilink
        arrow-up
        14
        arrow-down
        1
        ·
        6 hours ago

        It’s not possible to eliminate IP rights under capitalism.

        It is. Independent artists already gain basically nothing from IP law, because they can’t afford to defend their claims in court. The fact that their art (usually, before AI) isn’t stolen is a testament to the fact that there’s just not much money in doing so, while the fact that their art got stolen once AI hit the scene shows that the law doesn’t actually stop the rich from doing this stuff. Press the socialism button anyway, but in any case it’s mostly corporations who benefit from IP law.

  • deeprlyeh@lemmus.org
    link
    fedilink
    English
    arrow-up
    68
    ·
    9 hours ago

    That is awesome, anyone can train an AI model and use copyrighted materials. So this is the end of IP laws?

      • isleepinahammock@lemmy.blahaj.zone
        link
        fedilink
        English
        arrow-up
        14
        ·
        8 hours ago

        If anyone ever does try to argue this in court, they’ll just change the justification to “national security.”

        We need the most powerful AI we can make because our rival nations are building their most powerful AIs. Our rivals are using their AI to control drone swarms, perform automated intelligence-gathering, do mass automated hacking, and perform influence operations. Wars are quickly devolving into battles of drone swarms vs drone swarms, and the limit is on the people needed to control the. Our rivals are revolutionizing their supply chains with AI. They may even be performing vast automated influence operations designed to target individual citizens. In the near future, if you’re an American citizen, the Chinese government will know who you are personally, have a file on you, and will be using AI to try and psychologically manipulate you. We need to be able to fight this! We simply have no choice but to keep up. National security demands it. Companies participating in national security-crucial AI development will be exempt from copyright laws.

        I’m no lawyer, but a law to allow this almost certainly already exists. US law is full of national security exemptions. For example, I’m pretty sure in wartime IP laws can be waived. If you’ve invented some amazing new technology with military applications? If the government comes knocking and you just refuse to sell to them or demand a completely impossible price, there has to be a mechanism in IP law for those rights to simply be waived, or the rights taken by a form of eminent domain. There’s probably some provision of copyright law that would allow a president to void certain copyrights in a national security emergency.

        • Kirp123@lemmy.world
          link
          fedilink
          English
          arrow-up
          10
          ·
          8 hours ago

          Next time they send a cease and desist letter to some website hosting copyrighted content the website should just say they’re using the content to train AI models.

          • thallamabond@lemmy.world
            link
            fedilink
            English
            arrow-up
            2
            ·
            4 hours ago

            Good effort. You have been deemed a threat to ‘National Security’! Only the biggest ai firms get to do that.

    • NOT_RICK@lemmy.world
      link
      fedilink
      English
      arrow-up
      8
      ·
      8 hours ago

      You and I are unfortunately in the class of people that the law binds but does not protect

    • Grimy@lemmy.world
      link
      fedilink
      English
      arrow-up
      3
      ·
      edit-2
      3 hours ago

      Training a model with copyright material = fine

      Creating copyright material and distributing it = not okay

      I can infringe copyrights with Photoshop, or just copy pasting the image itself.

      We don’t need to knee cap tools and force everything behind subscription services with censorship just to satisfy the copyright industries craving for blood.

    • Korkki@lemmy.ml
      link
      fedilink
      English
      arrow-up
      2
      arrow-down
      1
      ·
      edit-2
      8 hours ago

      No, just another example that laws are for peyon proletarian poorfags. To the actual citizenry (large corporations) they are more of a suggestion or a guideline. US federal government just works to shaft the media and copyright holder megacorps in favour of the Ai megacorps, because rn the needs of AI megacorps are seens as more critical to the wider global future prospects of the community of corporate entities and global capital, which the American government represents.

  • eicker@lemmy.worldOP
    link
    fedilink
    English
    arrow-up
    107
    arrow-down
    1
    ·
    10 hours ago

    Trump’s government siding with OpenAI is peak MAGA economics: copyright is sacred until Silicon Valley needs millions of other people’s books, articles and art for free. Then suddenly property rights become an annoying obstacle to »winning the AI race.« Apparently America First now means AI companies first, creators somewhere near last.

    • driving_crooner@lemmy.eco.br
      link
      fedilink
      English
      arrow-up
      48
      ·
      9 hours ago

      Property rights are sacred until they need land to build a data center, then is eminent domain on the name of “public good”

    • bcgm3@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      5 hours ago

      Alrighty then, enjoy the end of creative people creating things, let’s all find out together how this ends 😐

    • Sims@lemmy.ml
      link
      fedilink
      English
      arrow-up
      4
      ·
      8 hours ago

      ‘Maga economics’ is just ordinary US Capitalist/Oligarch economics with panicking Oligarchs ! Did you really think that ‘your’ side of the super right-wing US uni-party would not shit on ANY principles as they see fit ?? Come on man It’s a big club… and you ain’t in it !

      Both sides are managed. It’s called ‘controlled opposition’. None of your ‘lefty’ values are officially cared about or listened too and you never had ‘Democracy’ in the US, so the Epsteins were always in control !! All of your opinions are nothing but thin air to your Oligarchy, but… somehow ‘Maga’ are bad, and you are good - without ANY of you having ANY democratic power at all !! Without ANY of you having good information, without ANY of you doing dilligent critical evaluations of your system. How incredibly captured are you people ? Do anyone of you even understand how US is run, or what actually happens in the world ?!?

      ANY agreement with ANY of your oligarch politicians/Epstein shtters are direct support for their insane behavior at home, and in their insane endless wars against the rest of the world, and guess what - both sides of the uniparty have had the same foreign policy - for 75 years !!! What you see now is nothing but Oligarch panic, and have nothing to do with any opinions of the Merican people - you are NOT in control, and you are NOT in the club !!

      Listen to George Carlin for f’s sake ! You are ALL - Maga’s and Dems alike - ‘controlled opposition’, 2 groups of manipulated, mentally raped, mind-colonised right-wingers that seem to believe whatever you are told by the most insane pschopathic elite, by the most insane media outlets that the Oligarchs actually OWN, and hence you live your entire life in the most ridiculous mis-information-sphere in the known world…

      You are ALL being played by the Epsteins pschos from the day you were born until now ! Wake up, Stop fighting each other on their command, Stop listening to their media, Stop hating who you are told to hate, Stop agreeing with their violent aggressions around the world - just S T O P ! !

  • Not_mikey@lemmy.dbzer0.com
    link
    fedilink
    English
    arrow-up
    2
    ·
    4 hours ago

    Can you reasonably train a local model? Inference is one thing and can be run on a decent gaming graphics card but training requires a lot more compute.

    • Dran@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      34 minutes ago

      You can absolutely train a nontrivial task model on a gaming gpu (image classifier, sound classifier, text model with a very structured input and output, etc). You can also post-train addon layers on top of existing open-weight models (LORA).

    • frongt@lemmy.zip
      link
      fedilink
      English
      arrow-up
      2
      ·
      4 hours ago

      Depends on the size of the training data, how big the resulting model is, and how fast you want it to finish.

  • mechoman444@lemmy.world
    link
    fedilink
    English
    arrow-up
    4
    arrow-down
    5
    ·
    edit-2
    2 hours ago

    So hear me out.

    From a purely strategic standpoint, the ultimate goal is artificial general intelligence. Eventually, humanity is going to get there. The question is who gets there first, and what they do with it.

    The country that achieves genuinely transformative AGI first could potentially gain an enormous strategic advantage. And if AGI is capable of significantly accelerating AI research itself, that advantage could compound very quickly: AGI helps develop better AI, which accelerates scientific and technological development, which produces even more capable AI.

    Whoever gets there second may not necessarily be permanently incapable of catching up, but they could find themselves facing a massive technological and economic gap that becomes increasingly difficult to close.

    And here’s where I think things get uncomfortable.

    The countries competing to develop this technology are not necessarily going to have the same restrictions. Some countries may be considerably less concerned about copyright, training data, ethical restrictions, censorship, or what subjects an AI is permitted to discuss.

    I’m not saying those concerns aren’t legitimate. They absolutely are. But from a purely geopolitical standpoint, there is a potential danger in one country imposing significant self-restraints while its competitors don’t.

    Because if AGI really does become as transformative as many people believe it could be, this isn’t just another technology race. It could become a race over scientific advancement, industrial capacity, military technology, economic productivity, and potentially the ability to develop even more advanced AI.

    And, honestly, I would rather the United States and other democratic countries reach that point first than China, North Korea, or another highly authoritarian state.

    That doesn’t mean I think democratic countries are incapable of abusing powerful technology. Of course they are. The point is that who controls a technology this powerful matters enormously.

    So when people talk about AI development purely in terms of corporate profits or whether AI companies should be allowed to train on particular datasets, I think they’re sometimes missing the much larger geopolitical picture.

    We’re potentially talking about a technology that could fundamentally alter the balance of power between nations.

    That’s a very different conversation.

    Disclaimer: I’m not necessarily advocating for AI or LLMs here. I’m simply trying to look at the situation from a purely strategic/geopolitical perspective and consider how governments and other actors might view the race toward AGI.

    • Zedd_Prophecy@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      6 hours ago

      I think they have every right to game the bubble because we would be doing the exact same if the situation were reversed. Bummer it’s all going to pop and we will all be using locals soon. They WILL keep building data centers though as long as they get away with it because these boomers and rich idiots want every ounce of our personal info.

  • Jo Miran@lemmy.ml
    link
    fedilink
    English
    arrow-up
    15
    arrow-down
    1
    ·
    edit-2
    8 hours ago

    THIS JUST IN!:
    Anna’s Archive to be renamed AI Training Repository of America.

    The lawsuits will drop, right?

  • Scrubbles@poptalk.scrubbles.tech
    link
    fedilink
    English
    arrow-up
    26
    arrow-down
    1
    ·
    10 hours ago

    From me just looking at it, everything they’re trying to do to force US AI supremacy seems to just generate and cause even more pushback. Ban cards to China? China’s ramping up GPU supplying and separating from US dominated supply chains. Ban model exports? They’ll make cheaper and easier models. Idk, whenever you force the market it usually backfires in ways you can’t expect.

    • Telorand@reddthat.com
      link
      fedilink
      English
      arrow-up
      27
      ·
      9 hours ago

      The thing is, they’re trying to force the market by actual force rather than slow and methodical coercion. US hegemony is over, because the buffoon in chief, who is terrible at domestic business, is also bad at foreign business. Shocking, I know.

      • Sims@lemmy.ml
        link
        fedilink
        English
        arrow-up
        2
        arrow-down
        1
        ·
        8 hours ago

        US hegemony is over

        That happened MANY years ago, and US elites came up with a plan to stop China by going through a shitload of wars and psy-op regime-changes - so called ‘Color-revolutions’ created by the CIA etc. All general moves have been outlined in think tank reports. What has happened that are different from normal, is that China suddenly accelerated and ran away from them, and in their panic they have had to trigger all their nasty attacks against all nations that in any way support China. Biden were the same nasty cnt as Trump, he just had only ONE small job for the oligarchs - Ukraine, whereas Trump has gotten them all, S. America, middelEast Iran, and oc the bad bad China. US are trying to regime-changing nations as we speak - exactly the same as under Biden, just much much faster…

        All of this is easy to figure out, but the western population was driven into a lynch-mob, an echo-chamber in the same way other nations were regime-changed. So, everyone is lied to, angry af, and as planned, doesn’t listen to counter arguments. ‘Putin-lover’, ‘Commie’ and other childish reactions are the norm, instead of actually investigating the ridiculous Claims from your elite.

        Yes, Trump is a buffoon, but honestly, can any of you actually name a single US oligarch politician that are NOT a buffoon ? Come on man, move on from this manufactured infighting and move away from US media. You are being manipulated to hate each other instead of the rich pscho cnts that actually RUN your country, and are actually harming millions of people around the world in a sick pursuit of more power !!

        • AlteredEgo@lemmy.ml
          link
          fedilink
          English
          arrow-up
          1
          ·
          5 hours ago

          That happened MANY years ago, and US elites came up with a plan to stop China by going through a shitload of wars and psy-op regime-changes - so called ‘Color-revolutions’ created by the CIA etc.

          I’m just curious - you mean something like “The Grand Chessboard - American Primacy and Its Geostrategic Imperatives” (1997)? I haven’t read that yet.

          But afaik they just fucking slept on China and that few of the regime change done in the last few 3 decades were aimed at it. Like they helped it get into the WTO and somehow they thought it would just become another neoliberal country that can be bought and controlled.

          I clearly see how the “unipolar moment” led to regime change aimed at the middle east and eastward EU and nato expansion of their vassal states, but I don’t see how that affects China.

          In any case yeah, China ran away and it’s pretty much over for US global supremacy, especially as they double down on not investing into their people: “In total, China now leads in 66 of the 74 technologies tracked, with the United States leading in the remaining eight” https://www.aspistrategist.org.au/aspis-critical-technology-tracker-2025-updates-and-10-new-technologies/

    • ✺roguetrick✺@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      arrow-down
      1
      ·
      9 hours ago

      Like you can force the market with good industrial policy of subsidies and oversight pretty effectively. See East Asia in general. Problem comes when those you gave power with those actions start directing your government to do stupid things for their benefit and they capture the very oversight that made it work in the first place.

      • willington@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        1
        ·
        edit-2
        4 hours ago

        It’s only a problem if you want to set up a good policy once and never touch it again.

        Life doesn’t work like that. “The problem with cleaning my room is that it gets dirty again.” Well, duh.

        Politics is a never ending process of vigilance and fighting. A policy should not be judged as “bad” simply because it requires vigilance and teeth at all times. Literally all good policies require now and will continue to require participation.