in the past few months i’ve given much more weight to chains of reasoning that are fuzzy or feel non-concrete. ie, the following now count as valid arguments

• that feels wrong intuitively
• x gives y energy (this feels like that)
• a is to b as c is to d, there must be something here

before, i wouldn’t buy it because it seems easily faliable. and it is, but its often better than nothing (if you have no ego, #writable).

in fact, refusing to reason by vibes may be stiffling. being comfortable working with these reasoning links allows you to 1. reason farther and 2. work with these partially-fleshd-out connections until they become fleshed out (it gives them a place in your mind before being codified)

thus, i think knowledge graphs should be transiently generated

 › thoughts