Measuring Epistemic Rationality

post by edoarad · 2019-12-28T18:46:24.041Z · LW · GW · 1 comments

Collecting some previous discussions on the matter, trying to reach a conclusion that would benefit me in my path to be more rational. Written with haste to serve as a reference for myself, possibly useful for you, and to be more accountable to be methodological in my rationality self-improvement.

1. The Rationality Checklist from CFAR has many rationality habits listed. I can go through it once in a while to feel progress, or focus on one that I feel I need to improve more and incorporate to my daily routine. (say- " I notice my mind flinching away from a thought; and when I notice, I flag that area as requiring more deliberate exploration. "). I'm not sure that I trust myself to reflect accurately. This is also "Not meant to be a way to get a ‘how rational are you?’ score, but, rather, to help you notice specific habits you might want to develop. "

2. Aliveness in Training [LW · GW] - Gives the analogy of a Martial Artist, that improves by working against an opponent. I have three thoughts -

i. Work with a mentor, seriously consider criticism from friends, write down more of my beliefs publicly.

ii. Measure my research success. (Correlated with lots of other things, and unclear, but that's actually what I care about)

iii. Make lot's of predictions

3. In Swimming Upstream [LW · GW], TurnTrout gives his personal experience with instrumental rationality. Parts of this story makes me think that I can measure how seriously I take my beliefs by collecting my own decisions and counting how many of them were due to newly generated beliefs. Not sure about that, because I may be biased in the wrong way here. I think that a fine idea here is to track all decisions made, and their reasons, and afterwards I'll understand how to analyze them.

4. It turns out that there is a Rationality Quotient, by Keith Stanovich [LW · GW]. I did not find the test, but I feel aversion toward it as a method to improve rationality. I think it's because I visualize myself Goodharting to it. It might be very useful as a one-time diagnostic tool.

In total, I think that I should

1 comments

Comments sorted by top scores.

comment by Gordon Seidoh Worley (gworley) · 2019-12-29T17:33:26.628Z · LW(p) · GW(p)

I'm not sure we understand enough to be able to get this right, sort of like we don't understand intelligence enough to get measurements of it right, but it seems worth trying since we might learn some interesting things along the way.