@pluralistic There is also the option - which is treated as being unthinkable, not even acknowledged publicly - of not allowing them to do the thing without research and having a regulatory framework built before it can cause catastrophic problems.
Like, with the LLMs - we could have done safety research, and learned they can lead to psychosis-like symptoms, before letting them loose on the general population, and regulated accordingly.
Regulation doesn't have to come *after* harm.

Hey look at this

* Every Drink in “Casablanca” (1942) https://bruces.medium.com/every-drink-in-casablanca-1942-348e7c543810

* clbre is a fork of calibre with the aim of stripping out the AI integration https://github.com/grimthorpe/clbre

* EU Report Distills AI-Training Lessons from Napster Piracy Era: Don’t Sue, License https://torrentfreak.com/eu-report-distills-ai-training-lessons-from-napster-piracy-era-dont-sue-license/

* Rebuilding Imaginary Futures: Il Versificatore, 2025 https://bruces.medium.com/rebuilding-imaginary-futures-il-versificatore-2025-3178a12be2aa

* John Varley, 1947-2025 https://floggingbabel.blogspot.com/2025/12/john-varley-1947-2025.html

4/