• 0 Posts
  • 54 Comments
Joined 2 years ago
cake
Cake day: July 1st, 2023

help-circle
rss













  • What resonated with me is people calling LLMs and Stable Diffusion “copyright laundering”. If copyright ever swung in AI’s favor it would be super easy to train an AI on stuff you want to steal, add in some generic training, and now you have a “new” piece of art.

    LLMs and Stable Diffusion are just compression algorithms for abstract patterns, only one level above data.




  • I think the worst thing about a Mary Sue is when their success comes trivially or randomly.

    What usually helps me is making the obstacle more specific and diving into those specifics when they’re problem solving. You’ll find most things we broadly group into large lumps, like martial arts, swordfighting, researching, medicine, ect. often have an overwhelming amount of details that not only separates good from bad, but also have specific dynamics that change depending on circumstances.

    If you want to make the successes feel earned, include enough detail about the problem that you can tell a story with the challenges involved. If your focus is swordfighting convey the kinds of techniques your protagonist know then put them up against opponents that can counter those techniques so they have to learn. If you focus is a doctor then instead of seeking out the Medicine Flower™, try conveying the roadmap to making medicine to the audience then make a story out of the process.

    I feel like Breaking Bad is a good example of this. It depends a lot on actual chemistry and every chemistry advancement is a plot point. Mainly it’s figuring out how to procure the ingredients and equipment without leaving evidence to get caught from.