@jeffjarvis we already have prefectly good API for news. RSS. I bet LLM's read RSS very easily, their training data must have a lot of examples.

#RSS #AI #LLM @davew

@jeffjarvis It's a realistic but depressing picture. AIs are even more opaque and game-able than either Search or Feeds were. What happens when we are all trained to blindly trust the computer? It will be even easier to manufacture propaganda through bias and omission.

@jeffjarvis I read you note - thanks for writing it.

One thing did leap out at me - you mentioned "transformative use". Transformative use remains a doctrine, especially after the Warhol case, that is not well anchored in our copyright statutes (I say "statutes" to reflect the text of our laws as enacted by Congress as opposed to the opinions of various judges.) (I remain unconvinced that "transformative use" has a statutory basis, but I do feel that our concepts of "fair use" do need to be brought into the 21st century.)

I do feel that there is something we have not yet pulled from our emotional dissonance between how we think of an AI "learning" a topic by reading the works of others versus a human doing the same thing (the latter being a well accepted practice.)

I, personally, still feel "AI" lacks the "I" part - an AI trained on Bach probably won't be able to come up with Coltrane. There's an aspect of creativity that seems to be lacking.