• 67 Posts
  • 597 Comments
Joined 2 years ago
cake
Cake day: December 18th, 2023

help-circle





  • YouTube and similar platforms are major content distributors. Content owners would not be able to enforce their rights without automated surveillance, whether through Content ID or another system. Meanwhile, it is generally not economical for small creators, influencers, to litigate erroneous detections. So in many cases, Content ID is the ultimate arbiter of copyright.

    When the likes of Taylor Swift or Donald Trump want to collect their licensing fees, they will use an automated system. And if some hapless citizen feels harassed by fakes and wants to make them harder to find, that will what be their only choice.


  • No surprise there.

    People seem to have forgotten that liberal democracy prevailed not through moral arguments or appeals to conscience. It prevailed because it was stronger in war and in peace. It makes more and better weapons when it has to, and enables its citizens a better life otherwise.

    Look how Hungary oppresses LGBT people. That costs a lot of money and eventually only means that members of that minority pay less taxes and contribute less to the economy.

    Look to the US. The poorest states in the nation are those that were committed to racial oppression. The US would be even richer without that frivolous waste of human potential. And not just the victims but everyone.


  • It would work like Copyright currently works.

    Yes, exactly. Content ID is a major part of how copyright currently works. The content industry convinced US courts that merely reacting to take-down notices was not enough. Companies hosting user generated content need to proactively search for infringing content.

    In the EU, written law goes somewhat further. In both the US and EU, this explicitly does not require a lawsuit. It is an automated process for most practical purposes.

    I can’t predict how Danish courts will see this. There are currently cases ongoing at the EU level that will make things clearer in that respect.














  • Not comparable.

    Samples are actual copies which are part of a song. Someone might claim that a hip hop artist just steals the good bits of other people’s songs and mashes them together without contributing any meaningful creativity on their own. Well. History shows that such arguments were quite foolish. Nevertheless, the copies are there, and they do add value to the new song.

    To get an LLM model to spit out training data takes careful manipulation by the user. This rarely happens by accident. It also does not add value to the model. It does the opposite: The possibility of accidentally violating copyright lowers the value.