In last week’s note, I wrote about how I use AI to find recurring patterns in my daily notes.
The idea came from Donella Meadows’ work on feedback loops.
The next question that came to mind was: What happens after recognising a pattern?
Naturally, I look for a solution to “fix” or improve the problem. I adjust my routine, change the process, or try something different.
But Meadows’ work on delays made me wonder:
Am I giving the last adjustment enough time to do its work before making another one?
I’ve been experiencing midnight wake-ups for some time. This has been showing up in my Pattern Tracker reports.
Recently, I started practising meditation and breathing exercises, hoping they might help.
After two nights, I didn’t notice much improvement. I started questioning whether the exercises were working and looking for other solutions.
In Thinking in Systems, Meadows used a car dealership to explain how delays affect a system. When invetory runs low, the dealer orders more cars. But before those cars arrive, the dealer may order even more.
By the time the deliveries arrive, there are too many cars.
The dealer made another correction without fully accounting for the previous one.
That made me wonder whether I was judging my own practice too early.
What exactly was I expecting from it? To feel calmer when I wake up? To fall asleep again more easily? Or to stop waking up in the first place?
These are different outcomes.
Two nights might tell me whether an exercise feels relaxing. But are they enough to decide whether it can improve my sleep?
I don’t know yet. At least not after two nights.
And unlike the cars already on their way, I don’t know whether the breathing exercises will eventually work.
The absense of an immediate result doesn’t necessarily mean that an adjustment has failed.
Making another adjustment doesn’t necessarily make things better.
Maybe a useful question we could ask is:
What result am I expecting, how long should it reasonably take, and what would make me reconsider sooner?
I can continue observing without immediately changing what I’m doing.
Not every observation needs to become an action.
Sometimes, the previous action is still working its way through the system.
Until next time,
Gav


