In my recovery from burn-out, I’m balancing my energy, learning to take the right amount of rest and breaks, and better follow my energy when doing work or chores. I still find it hard to get started on most of anything, and get myself going.

I’ve taking up interstitial journaling to help me switch between tasks (in which you describe what you just did, any thoughts or feelings about it, what you think you might do next and if there’s anything that makes you balk at it etc.), and that has been very useful in the past year or more. It helps me ease into the next thing.

Starting is however a different thing. I find I need to talk myself through how to choose between different possibilities, and also to make tasks smaller. To do that it’s currently not enough to think about by immediatel making lists or outlines of the steps / components involved as I normally would and then move them around or split them up.

What helps is talking myself through things by formulating steps in long form, not just list them. That yields formulations of the type "ah wait I better do this first then and afterwards go do that, but wait I also need to think of somesuch, and perhaps make a call". Making internal dialogue explicit. It feels like reading along with texts that LLMs generate for themselves to consume in my AI sandboxes. Where there is some very verbose output going back and forth for some time, and then at the end it gets collapsed into a 5 point todo-list.

Alsmost as if I am prompt-engineering myself to get the responses and output I want. Not anthropomorphising AI, but the other way around, basically chat-botting my self. Not at all sure what to think of that! If I do it in speech-to-text (Spokenly running a local model), making such inner dialogue explicit, it is even a lot more words for what basically is a task list. Odd.