so much of my research now is voice inputting several minutes of thoughts about potential new ideas into ai and then either halfway through the input I realize the next step or the model spits out a bag of words one of which clicks me into gear

once it had a legitimately good idea fully formed that I had to take a few minutes before I understood. But usually it’s non-sense (or I’m just missing its brilliance) that happens to be close enough to sense that it’s useful for brainstorming