Does your theory apply to the content consumed during the training of LLMs? If not why not? Where should we draw the line regarding intellectual property rights?
2. Information and knowledge is for all.
3. The line to draw is 10 years to commercialize, and then release into public domain. Statute of Anne was 14 years (https://en.wikipedia.org/wiki/Statute_of_Anne), but 10 is better for today's age.
I'd vote for politicians who push for laws like this. Sadly none do.
Even so, I think in most cases if a piece of software has been unedited for 10 years it’s either feature-complete or obsolete. If it’s been under development for those 10 years, the original version being released into the public domain probably isn’t a significant threat to innovation by way of closed source improvements being made.
The issue here isnt the IA's provision of brand new books that are still being published; this few would say should be legal. We're talking dead authors, books no longer in print, or books published so long ago that the second-hand market (which offers no pay to the author) is the place to find them.
As soon as works transition to "second-hand markets" we're no longer talking about the labour of the author being remunerated. At this point, it's pretty clear that it's a net benefit to society to make creative works publically available.
2. LLMs are an incredible step forward in humanity's progression.
3. 10 years from publishing, then public domain afterwards. The content must be commercialised to have copyright; if it's available for free, it should already be public domain, because copyright is supposed to help you make money, not help you control the use of information.