Michael Nielsen: “In the Mission in SF there was a flyer posted on a street pole, with a QR code that would let you download the Llama weights”

One of the biggest challenges I face when talking to “more, better regulation, please”-type Civil Society is to explain that:

  • All the <whatever> restrictions they demand in order to rein-in BigTechPower™ will also be placed upon academics, journalists, hobbyists, activists, and other PeopleLikeThem™
  • When that happens they will have to start a counter-campaign to demand RegulationYesButNotLikeThat™ and that campaign will fail if/when security services feel that “normal people” should remain arrestable for performing unregulated <whatever>
  • Because <whatever> is software and therefore speech (protected, free, or otherwise) there is literally nothing which can be done to actually restrict the global rise of <whatever>
  • Michael A Nielsen put it nicely in an observation regarding the leak of Meta’s LLaMa AI weights:

    Open source is forever. In the Mission in SF there was a flyer posted on a street pole, with a QR code that would let you download the Llama weights (this was when Meta was still ambivalent over the release)

    https://www.threads.net/@michael.a.nielsen/post/Czg51oaPgpH/

    In the naive old days we used to say “information wants to be free”, but actually it’s more like “learning and speech will be shared, in spite of everything done to prevent it”

    #artificialIntelligence #essay #informationWantsToBeFree #llama #meta #michaelNielsen

    https://alecmuffett.com/article/108280

    Meta’s powerful AI language model has leaked online — what happens now?

    Meta created its new LLaMA AI language model to further research into problems that affect chatbots like ChatGPT and Bing. But a week after it was announced, the model was leaked on 4chan, prompting worries about misuse.

    The Verge