Kids under five anecdotally have a harder time learning to count, learning that there's still the same number of balls after you re-arrange them than before (many pre-schoolers will count them all over again), etc. So a math program for pre-schoolers needs to be built from smaller building-blocks than a program for adults. I imagine it's similar with rationality.

That needn't mean teaching math or rationality to pre-schoolers is silly. My mom played a lot of math games with me in the zero to five years, and while it's hard to tell effects without larger sample-sizes, and despite the Judith Rich Harris results, it seems plausible that this gave me math farther into my bones than I would otherwise have gotten it. (E.g., a large fraction of my pre-age-eight memories involve mathematical concepts, usually ones I was just sponaneously thinking about.)

As to the feeling of safety, there're a lot of early rationalist thinking skills that can be taught without eroding the kid's trust in their parents and stable home. Math, noticing how you're feeling and what effect that has on your actions, asking why gadgets work and enjoying the exploration, turning questions into testable hypotheses...