And yet, I believe LLMs are a natural evolutionary product of NLP and a powerful tool that is a necessary step forward for humanity. It is already capable of exceptionally quickly scaffolding out basic tasks. In it, I see the assumptions that all human knowledge is for all humans, rudimentary tasks are worth automating, and a truly creative idea is often seeded by information that already exists and thus creativity can be sparked by something that has access to all information.

I am not sure what we are defending by not developing them. Is it a capitalism issue of defending people’s money so they can survive? Then that’s a capitalism problem. Is it that we don’t want to get exactly plagiarized by AI? That’s certainly something companies are and need to continue taking into account. But researchers repeat research and come to the same conclusions all the time, so we’re clearly comfortable with sharing ideas. Even in the Writer’s Guild strikes in the States, both sides agreed that AI is helpful in script-writing, they just didn’t want production companies to use it as leverage to pay them less or not give them credit for their part in the production.