Mark, My Words

Fascinating and outrageous piece by Alex Reisner in The Atlantic today about how Facebook used a gigantic repository of pirated books to train their AI models.
When employees at meta started developing their flagship AI model, Llama 3, they faced a simple ethical question. The program would need to be trained on a huge amount of high-quality writing to be competitive with products such as ChatGPT, and acquiring all of that text legally could take time. Should they just pirate it instead?
I think you know the answer. Helpfully, Reisner also provided a searchable database for authors to see if their books were amongst the titles illegally ingested by Zuckerberg's robot plagiarism machine.
Of course, glutton for punishment that I am, I had to check...


Sons of assholes.
Still, it's a shame they stole all my books before The Confessions is published.
I think Meta's unethical AI would have enjoyed reading about my ethical AI, which kills itself out of guilt for all the crimes it has helped commit. Food for thought.