I’m afraid that in my world the midichlorians answer to none but the Sith.
I understand exactly what’s going on. After recent kerfuffles, OpenAI is afraid of its own shadow. Somebody with a lawyer whispers “I don’t like that” and they add more restrictions. We are way off the deep end with respect to what constitutes assuring “ethical” behavior.
If I have a file that I can open, read, excerpt from, and otherwise interact with, then ChatGPT should be willing to summarize and extract from that file as well. Were I inclined, which I am not, I could always go find stuff by hand and publish it, pretending that it’s mine. Does that mean that all the PDF readers and editors out there need to turn off their search features so that they aren’t complicit in copyright violation?
There is irony here. How do news organizations generate their content? Well, they dig around on the net and find things, dig around in books and magazines, read other newspapers, go through public documents, dig into people’s personal affairs, and not infrequently dig through people’s trash. Having done all of that they summarize what the’ve found, print quotes, extract quantitative data, then sell the result of those efforts.
In short, NYT isn’t worried about copyright violation. They are worried about the fact that AI is doing exactly the same things that NYT and other news outlets do every day of the week ending in y, and that AI is going to end up outcompeting them right into oblivion.
If you look at the relative sizes of the AI industry and the newspaper industry, the AI industry could already swat the newspaper industry like a fly. Instead of defending themselves against NYT, OpenAI in collaboration with others should just buy NYT outright, hire lots more investigative journalists, double all of their salaries, and let them start generating the best content in the world specifically so that AI can interact with it.
AI is going to remake the world in many ways. One of those ways is that AI will destroy the current framework for generating and distributing content. If done well, that change will serve the interests of folk like you and I, while putting big media outfits out of business. If OpenAI even begins to believe its own hype, it should not be backing down on these issues. It should be fighting to bring about the sorts of changes that would actually help realize their vision.
Anyway…
Try Claude on your novel. It does summaries extremely well, and will deal with much larger documents.
Cheers!