I strongly second this. Even just standing up and walking away from my desk for a few minutes is often enough to trigger insight. And sometimes I'll take a longer break from something like a few days or even weeks and then come back to it with a fresh perspective.
On a side note, I've noticed this is particularly useful when dealing with agents too. Claude would happily bore to the center of the Earth if I let it, but taking a step back and reframing the basics of the problem often leads to a simpler and stronger outcome.
I agree, "do something" is the way out of analysis paralysis.
That said, I also chat with my AI a LOT now, and it helps me clarify things. Yes, there are risks and caveats, no I don't let it make my decisions, but it helps sharpen them a lot. And it plans my day for me, which is great, takes away decision fatigue and if I disagree I just tell it and it does it differently. I even let it track how long I spend on tasks recently, also good for preventing rabbit holes.
And yes, I'm also very curious about many things, and am great at rabbit holes. But a clear goal can help you prevent unnecessary rabbit holes (leaving enough time for the occasional very enjoyable rabbit hole, something we shouldn't lose!).
I usually just check if I can undo it. If I can, I go with whatever seems ok and fix it later, and that's most decisions. I only really dig into the ones I can't undo.
Go for a walk.
That encourages the incubation phase of the creative process. Where your subconcious defrags your thoughts and makes things clear again. Search for 'Henri Poincaré bus' for a tech/maths example.
I strongly second this. Even just standing up and walking away from my desk for a few minutes is often enough to trigger insight. And sometimes I'll take a longer break from something like a few days or even weeks and then come back to it with a fresh perspective.
On a side note, I've noticed this is particularly useful when dealing with agents too. Claude would happily bore to the center of the Earth if I let it, but taking a step back and reframing the basics of the problem often leads to a simpler and stronger outcome.
Notice that you get the most valuable information from your environment by attempting to do something.
I agree, "do something" is the way out of analysis paralysis.
That said, I also chat with my AI a LOT now, and it helps me clarify things. Yes, there are risks and caveats, no I don't let it make my decisions, but it helps sharpen them a lot. And it plans my day for me, which is great, takes away decision fatigue and if I disagree I just tell it and it does it differently. I even let it track how long I spend on tasks recently, also good for preventing rabbit holes.
And yes, I'm also very curious about many things, and am great at rabbit holes. But a clear goal can help you prevent unnecessary rabbit holes (leaving enough time for the occasional very enjoyable rabbit hole, something we shouldn't lose!).
Military has scouting. Engineering has spikes.
It sounds like your problem is you're afraid to get stuck in the rabbit hole. Go down the rabbit hole, but make sure you bring rope, i.e. come back.
Some techniques I've seen work:
Set a deadline, either real or artificial. Or find an external forcing function like a commitment to present on <topic>.
Determine if a decision is type 1 or type 2: https://www.theuncertaintyproject.org/tools/decision-types
Think back to other similar decisions you've made in the past and what the process and outcome was.
I usually just check if I can undo it. If I can, I go with whatever seems ok and fix it later, and that's most decisions. I only really dig into the ones I can't undo.