Mediocre LLM slop.

This is a solvable problem. Just make a LoRA of the Alice character. For modifications to the character, you might also need to make more LoRAs, but again totally doable. Then at runtime, you are just shuffling LoRAs when you need to generate.

You’re correct that it will struggle to give you exactly what you want because you need to have some “machine sympathy.” If you think in smaller steps and get the machine to do those smaller, more do-able steps, you can eventually accomplish the overall goal. It is the difference in asking a model to write a story versus asking it to first generate characters, a scenario, plot and then using that as context to write just a small part of the story. The first story will be bland and incoherent after awhile. The second, through better context control, will weave you a pretty consistent story.

These models are not magic (even though it feels like it). That they follow instructions at all is amazing, but they simply will not get the nuance of the overall picture and be able to accomplish it un-aided. If you think of them as natural language processors capable of simple, mechanical tasks and drive them mechanistically, you’ll get much better results.

Leading to either having to carefully double check what it suggests, or having fix bugs in code that I wrote but didn’t actually write.

100% this. Recent update from jetbrains turned on the AI shitcomplete (I guess my org decided to pay for it). Not only is it slow af, but in trying it, I discovered that I have to fight the suggestions because they are just wrong. And what is terrible is I know my coworkers will definitely use it and I’ll be stuck fixing their low-skill shit that is now riddled with subtle AI shitcomplete. The tools are simply not ready, and anyone that tells you they are, do not have the skill or experience to back up their assertion.