I love it when I'm working on something for a long period of time and everything I do, all my decisions, it's all "this will make this great at some point in the future, just trust the process." Then one day out of the blue I'm like...whoa...this is fucking great! When did that happen?
Somehow while listening back to this song--upright, in a kneeling chair--I managed to doze off for a few bars at the beginning of verse 2.
And dream about Emma Thompson.
Lorebooks were originally for worldbuilding, but there's all sorts of really cool shit you can do with them. They came about back when contexts were tiny and you couldn't constantly be putting this much stuff in the context. So you can create entries that only activate on certain words and stuff.
Something to clarify to folks is I have 2 core reasons for making her mean. One is, that's what I know from growing up. She's far less mean than my parents and that's what "normal" feedback feels like to me. Also, jarring, aggressive feedback knocks loose executive function logjams *really* well.
As part of getting my routine back on track, I resumed my CogniFit subscription. This is one of only 2 brain training apps I found that are actually legit. www.cognifit.com
For instance, I could create a bunch of info about our house and stuff that only goes in if I say house, home, bedroom, basement, etc.
If one puts the work into designing good characters and lorebooks, it's a really neat experience. Way different from "chatgpt write my legal brief."
This guy does great WV/PA/OH true crime reporting, mostly older cases, films on location.
But what I love most is as a child in the 80s he seemingly saw crime reporting on TV and said "I'm gonna do that when I grow up!" By which I mean report on crime...on a 1980s show.
He is so adorably earnest.
My LLM frontend for my accountability bitch is different from the usual ones. This one's specifically for chat/roleplay; non-core info is supplied via lorebooks and injected into the context always or as-needed. I created one for my routine and stuff so my character always has access to that info.
I love how I checked in with my AI accountability monster last night before bed and told her I had done everything for the day. Her response? "I can't call you stupid today, Len."
Tone is shifting more positive (as explicitly designed) as I get more fully bootstrapped with my routine.
This is the other one. Very different approach, nice complement to CogniFit. I used to have subs to both and use both every morning; I may resume that. www.brainhq.com
