GNU social JP
  • FAQ
  • Login
GNU social JPは日本のGNU socialサーバーです。
Usage/ToS/admin/test/Pleroma FE
  • Public

    • Public
    • Network
    • Groups
    • Featured
    • Popular
    • People

Conversation

Notices

  1. Embed this notice
    Koen Hufkens, PhD (koen_hufkens@mastodon.social)'s status on Tuesday, 10-Dec-2024 07:00:17 JST Koen Hufkens, PhD Koen Hufkens, PhD
    • Cory Doctorow

    Finished the "The Internet Con: How to Seize the Means of Computation " by @pluralistic. Necessary reading in this day and age, for a few reasons:

    1. It introduces the concept of universal computing devices really well
    2. and in this context how (internet) technology underpins societal and justice movements, stifled by software copyright "wrappers"

    In the context of the former it also shows were we are headed IMO.

    1/

    #privacy #copyright #software #opensource #enshitification #technology

    In conversation about 5 months ago from mastodon.social permalink
    • Embed this notice
      Koen Hufkens, PhD (koen_hufkens@mastodon.social)'s status on Tuesday, 10-Dec-2024 07:56:18 JST Koen Hufkens, PhD Koen Hufkens, PhD
      in reply to

      When turning the tables, what will interoperability look like in this new LLM software age. If it is at all possible.

      Will people be prosecuted for using LLMs "the wrong way", asking "bad" questions? Is prompt hacking, or prompt injections, really a hack and reason for criminal charges?

      What is considered a copyright infringement when this meta-software is built largely without consent?

      It leaves me with the feeling that LLMs are the new walled gardens and an interoperability disaster.

      3/

      In conversation about 5 months ago permalink
      Cory Doctorow and Paul Cantrell repeated this.
    • Embed this notice
      Koen Hufkens, PhD (koen_hufkens@mastodon.social)'s status on Tuesday, 10-Dec-2024 07:56:18 JST Koen Hufkens, PhD Koen Hufkens, PhD
      in reply to
      • Cory Doctorow

      It is Daniel Davies who popularized the idea that these LLMs are an excellent "accountability sink" (@pluralistic put it on my radar). Plausible deniability. Shroedinger's search engine, where things are correct and hallucinations at the same time. Computer said no.

      Where (poor) AI models were previously limited to certain industries, they are now injected in every application possible (i.e. blowing up that potential legal surface).

      Software ate the world once, LLMs will do it once more.

      fin/

      In conversation about 5 months ago permalink
    • Embed this notice
      Koen Hufkens, PhD (koen_hufkens@mastodon.social)'s status on Tuesday, 10-Dec-2024 07:56:19 JST Koen Hufkens, PhD Koen Hufkens, PhD
      in reply to

      When I look at the large language models (LLM) I see meta-software. Software which isn't. It is multi-modal, yet disembodied. It is search, but not search.

      Where this tech exposes a huge attack surface security wise, it opens up an equally large legal attack surface.

      Currently the focus has been on copyright infringements by LLM, but by and large the weight of the industry will judge in their favour and content will be licensed.

      2/

      In conversation about 5 months ago permalink
    • Embed this notice
      Koen Hufkens, PhD (koen_hufkens@mastodon.social)'s status on Tuesday, 10-Dec-2024 10:56:08 JST Koen Hufkens, PhD Koen Hufkens, PhD
      in reply to
      • Cory Doctorow
      • Hao Ye

      @hye @pluralistic It is the dog ate my homework excuse for software, e.g. the whole itch.io snafu. These things rarely end in favour of those most impacted. How do you file a CVE for an inherent stochastic system?

      In conversation about 5 months ago permalink

      Attachments

      1. No result found on File_thumbnail lookup.
        Download the latest indie games
        from @itchio
        itch.io is a simple way to find, download and distribute indie games online. Whether you're a developer looking to upload your game or just someone looking for something new to play itch.io has you covered.
      Cory Doctorow and Paul Cantrell repeated this.
    • Embed this notice
      SaftyKuma (saftykuma@mastodon.social)'s status on Tuesday, 10-Dec-2024 10:56:08 JST SaftyKuma SaftyKuma
      in reply to
      • Cory Doctorow
      • Hao Ye

      @koen_hufkens @hye @pluralistic

      I almost think that THIS is the biggest draw to LLMs for the corpo types. They can instruct their software to do shitty things, but then wave their hands and blame "an AI we can't possibly control" when they are called out.

      In conversation about 5 months ago permalink
    • Embed this notice
      Hao Ye (hye@glammr.us)'s status on Tuesday, 10-Dec-2024 10:56:10 JST Hao Ye Hao Ye
      in reply to
      • Cory Doctorow

      @koen_hufkens @pluralistic Hmm, when I think about it that way, chatgpt is a lot cheaper as an accountability sink than Bain / McKinsey / et al.!

      In conversation about 5 months ago permalink

Feeds

  • Activity Streams
  • RSS 2.0
  • Atom
  • Help
  • About
  • FAQ
  • TOS
  • Privacy
  • Source
  • Version
  • Contact

GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.

Creative Commons Attribution 3.0 All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.