Share.

    30 Comments

    1. Today’s large language models (LLMs) are developed using data from the public, meaning they draw on the collective knowledge and contributions of countless individuals. Given this foundation, it would be unethical for these technologies to be owned or controlled by a single entity. Instead, the management and governance of LLMs should reflect the collective nature of their creation. Socialism, which emphasizes communal ownership and equitable distribution of resources, offers a fair approach to overseeing and regulating this technology. By ensuring that LLMs remain a shared resource, we can promote greater accessibility, innovation, and ethical use, benefiting society as a whole.

    2. TheLastSamurai on

      Correct because corporations are either going to see us starved off and become serfs or get us all killed from their hubris

    3. Which flavour of socialism were you thinking, exactly?

      I presume you’re not rooting for the classic model where the “AI” is state-owned and the state also deploys electric fences, minefields and guards with attack dogs to prevent citizens from seeking alternatives.

    4. Elegant_Studio4374 on

      Direct democracy non representative socialism where all the work is done by robots.. fixed that for you.

    5. NeuroticKnight on

      FB’s Ilama, and Hugging Face have large opensource models, unfortunately they are the most at attack, since Google and Microsoft and Apple can pay for lawyers, but the Opensource community doesn’t, just

    6. It’d be cute if we reared the models on ethics and philosophy. Too bad we reared it on DeviantArt and Reddit, instead.

    7. At the end of the equation of solving people’s problems you arrive to socialism.

      If the AI consider every life is precious and deserve respect, it will arrive to socialim.

    8. BCDragon3000 on

      …and boom. there’s the buzzword that divides our world and suddenly the next 45 years of politics got so much more annoying

    9. I came to this conclusion the other week. Gen ai trained on public data to mimic copyrighted material should be made available as a nationalised or realistically, internationalised resource.

    10. noonemustknowmysecre on

      Yeah ok. Capitalists don’t really care too much about what’s ethical anyway, so this doesn’t change much. Like, it’s not even in their list of wants and desires. “Whatever makes the most money”. Now, if it would be massively costly to thwart the masses rioting and rebelling and murdering if things aren’t handled fairly, THEN they start to see the light.

      But the idea that these things are trained on everyone’s work, so everyone ought to benefit? Sure. Sounds good. I’d even like to go one further and simply track what is useful to the LLMs and who wrote it. They’re currently black boxes and the training successfully obfuscates what precisely is being used for what, we could develop ways to keep track. If, in the end, the weight of all of Reddit’s input is zero, then we’re not really contributing anything here to AI. It’s likely going to be pretty evenly and broadly distributed though.

    11. Lol, delusional to think that AI will be used for anything other than exploiting.

    12. Socialism (the collective ownership of the means of production) has never resulted in an ethical system before. Why would it happen with AI?

    13. KAHANEchai1947 on

      Except the coding, prompting, servers, chips, power, etc all came from capitalist private sources so that kind of puts a dent in your argument

    14. Tall-Dragonfly-2392 on

      Socialism never works, never will work, because people are inherently selfish, and those who seek to lead, are naturally egotistical and narcissistic (sp?). Putting those traits in a computer doesn’t make it any better.

    15. If an employer is allowed to fire employees and replace them with AI, employees should be allowed to develop their own AI to do their job for them and keep collecting their salary

    16. Motor_System_6171 on

      The AI Dividend. It was built in our data, so we all own a piece of every single token.

      Nontime like the present to start paying up.

    17. By that standard the same applies to the economic system in general. Don’t get me wrong, I actually agree, just not for those reasons. I agree because this is the way out of our current economic system – this moves us away from scarcity, but so much of our society is built on the assumed premise that we’re going to have to deal with fairly distributing the results (at least until we actually move fully past scarcity).

    18. ProofChampionship184 on

      The only ethical model for any type of human activity is socialism, because socialism means democracy and liberty. That is our dilemma as we approach the precipice: capitalism or freedom. It is a choice.

    19. MYDOGSMOKES5MEODMT on

      People who speak in absolutes are slowing tying their own ethical nooses

      We have *NO IDEA* what the right move about AI is yet, for an exceptional quantity of reasons

    20. If you Americans really want to become like North Korea, Cuba and Venezuela, maybe you can change places with millions of masochists who want to flee from of these socialist paradises.

    21. Socialism is giving the means of production to workers.
      That’s what it is.

      AI will replace human labor- all labor. Socialism, much like Capitalism, relies on labor. Every econo.y system has.

      We need a paradigm shift for an era without labor where all citizens share in AI and its fruits with some unheard of for of resource allocation- not Socialism.

      I do think we should advocate for UBO though- Universal Basic Ownership in any models which have been built using our general training data- as well as any derivatives.

    22. FaultElectrical4075 on

      Full labor automation is both an inevitable consequence and a possible killer of capitalism.

    23. If the AI considers every life precious with the same rights then it will not get to socialism but rather libertarianism.