Esme Povirk @madewokherd@computerfairi.es

So far ChatGPT seems to be pretty good at this.

I think maybe the reason I don't like debuggers (the kind you type commands into) is that it feels like an execution challenge, where I have to perform the correct sequence of commands and actions to get the information I want. With print statement debugging I just write more code and if it's not good enough I fix it.

So basically I want a debugger I can program ahead of time.

Finding that in difficult debugging situations where I'd previously give up, my response now is more often to improve my tools.

Hypothesis: One of the few things LLMs are good at is metaphor. Human minds respond well to metaphor and can be changed and observed via imagination (which doesn't have to be visual - anything you can experience is a possible way to imagine). The problem is often finding the right metaphor what you want.

Therefore, LLMs may be good for helping to observe and make changes in one's mind through metaphor.

I am either too tired to give my talk, or the perfect amount of tired to give my talk.

I have begun to think that any action that slows down oppression, even a little bit, matters. Any additional resources used means fewer available to oppress other people, and a delay to those plans. It doesn't have to change the winds to make a difference.

Planning a 27-hour train trip because flying (not the actual "being in the air" part but everything around it) sucks that much.

Being sober and going to bed at a reasonable time is underrated.

This morning I set a roll of toilet paper so it would hang down the side of my bathtub, so that a spider could use it to climb out. It was adorable to watch. I love spiders.

(Usually I try to grab them by holding something flat underneath them and then drop that outside the tub, but I thought this might be less scary for them? Plus, it can still help while I'm away.)

Bathtubs should have textured bits so spiders can climb out.

(While writing this, I programmed myself to do something a little silly later, with no hypnosis or trance around it. Mostly to observe what I was doing. Will be interesting to see if it works.)

I think that when someone looks at you very seriously and gives you some warning about a precaution to take, so you don't burn or electrocute yourself or something, and you know it's important and somehow you always seem to remember when the time comes, that's the same thing as programming a post-hypnotic response.