Post

Is there a link to the full paper?

No, this is for a talk I'm presenting at a conference next week in Rome, organized by @eryk.bsky.social et al: arthist.net/archive/52572 But I will try to polish and publish it afterward!

Why would that be surprising? LLMs - AI slop - is best known for being drab and predictable and non-creative.

What's surprising is that the giant trillion-parameter models are even more predictable than the tiny ones I can run on my own PC. That, plus the fact that 'aligned' or fine-tuned models are more predictable than their base model, suggests AI is optimized for predictability (a feature of slop?).

During training they are tasked to predict the next token and then get modified to predict the next token even better than before training. It's literally their job to be as predictable as possible! The bigger models have more parameters to do that even better.

an interesting wrinkle is that I strongly suspect Hemingway’s predictability increased with each draft: i.e., it’s a deliberate design decision

Definitely, no aesthetic judgment here per se. "Predictability" is part of its power.