Wonder if ChatGPT will be the last major model to be trained on the open web? robots.txt specifically disallowing crawling from LLMs unless getting paid for the data?

If AIs can generate enough value, it might be worth paying armies of Mechanical Turk-style workers to manually visit and rewrite web sites for copyright-approved training

Facts and ideas can't be copyrighted, only particular expression

Freelance data engineer | Backend and data engineering for Spotify, Candy Crush, helped build SF-based appsec startup | Building web3gains.com | saarw.twitter

Improving Shell packaging in the Nix ecosystem. Looking for the seams. I code/write/edit/read and give a fuck as a service. 

Journalists and academics have copied facts from encyclopedias forever to write their own articles

This would double down on the risk LLMs are laundering misinfo, no?

I also expect avoiding liability is more complex than just paying someone with a pulse to thesaurus words. I think a jury would agree the work was copied if there's deep, undeniable structural similarity.