Technology

It will look like a man playing with a dog, and we are not the man.

Okay, this is some useful feedback on what a couple people think in general about the idea of AI acting as a frame for human creativity. Thanks for that!

@[email protected] , would you consider trying the thing we made and telling us, specifically, what about this you don't like? I agree that the idea of a "text being queried again" is problematic, but that's not exactly what we're suggesting here - the text is, at least arguable, run through the subtext of a great many other texts.

So is there scope for seeing an encounter with the thing we made as a projection of a specific text through all the different texts used to train a generative model? Or is an individual reader's experience too important and special to ever claim that we can use a statistical "close-and-distant" reading to do something meaningful to something that's already part of the canon?

Please give it a try! Negative feedback is good, and we're also sceptical about this, but reference to a specific experience helps.

Human and AGI collaboration might be interesting, if ever real AI actually develops. But I wouldn’t call augmenting or probing of existing works of fiction with rehashed LLM sludge collaboration, I’d call it glorified and advanced plagiarism at worst, and low quality cliff notes at best.

I would much rather read a work of creative fiction from a human being than to encounter autocorrect word predictions written into paragraphs. The idea that the text itself can be queried to gain additional meaning divorced from the author’s intention strikes me as unrealistic and not faithful to the person who originally crafted the words.

Though I’m obviously biased against LLMs being used for this kind of thing, from lots of experience seeing how crappy they are.

OG horror..