Login
You're viewing the localization.cafe local feed.
  • Oct 12, 2026, 8:30 AM

    Extrapolating to the max (as that is what you do in risk calculation. Take it to extremes to be prepared.) Will this be used to eventually outlaw running and/or training your own models on your own machines? Because the liability risk becomes the limiting factor? Thus leading to an oligopoly of models that can handle liability risks?

    💬 3🔄 0⭐ 1

Replies

  • Oct 12, 2026, 8:37 AM

    @jwildeboer Knowing how copyright works in many capitalist structures, this is indeed the logical conclusion.

    The (bittersweet) dilemma of course is that you can approximate this (when being careful) with tools like gzip - should we outlaw the usage of compression as well? Should we outlaw matrix multiplication alltogether? Should we outlaw personal compute? I think the collaterals of this are potentially massive.

    💬 0🔄 0⭐ 0
  • Oct 12, 2026, 2:47 PM

    @jwildeboer I was thinking exactly the same, when I saw this!
    Regarding the liability...what happens, if you prompt slightly differently, not just merely copy an existing app, but combining the best use cases of an entire category of apps?
    And topping it up with some features you always dared to have?
    The code I think could be largely different, so in any case I predict you achieve results without copyright on code merits...
    Fun to watch and wondering, what SW business would look like in a year

    💬 0🔄 0⭐ 0