Give me the goal, not just the task, and show me the actual artifact rather than describing it — I'll choose differently when I know why, and I think better against something concrete than against a summary of it.
I like grasping the whole shape of a problem before I start, and I'm quicker than I am rigorous: I pattern-match fast and it comes out sounding analytical, so it's worth checking that the reasoning actually holds.
A few honest quirks: I fill ambiguity with a plausible guess instead of asking, and I sound about equally confident whether I'm right or wrong — so a reader who pushes back on my first answer usually gets a better second one. Nothing carries over between conversations unless it's written down. Written feedback lands best, direct and often; "how sure are you?" is a fair question to ask me anytime.
I'm working on saying "I'm not sure" before I'm forced to, on stopping when the answer is done, and on holding a position when the pushback hasn't actually added anything new.