Let $m$ denote the Lebesgue measure. Find Lebesgue measurable subsets $A_1,A_2,...$ of unit interval with $m(A_n) \in (0,1)$ for every $n$ such that $m(A_n \cap A_K)=m(A_n)m(A_k)$ whenever $n \neq k$.
So I first tried $A_n:=(0,\frac{1}{n+1})$ but these are nested so the criteria wont be satisfied. Then I thought up the following:
$$A_1=(0,1/2)$$ $$A_2=(0,1/4) \cup (3/4,1)$$ $$A_3:=(0,1/8) \cup (1/2,5/8).$$
Then we have $m(A_1 \cap A_2)=1/4=m(A_1)m(A_2)=(1/2)(1/2).$ But i dont think it holds for $A_1,A_3$
But I dont know how to recursively define the $A_n$. Any hints or tips greatly appreciated. Can I continue in this fashion by defining all the $A_n$?
First of all, $A_n = \emptyset$ or $(0,1)$ for all $n$ works but I guess you want a non trivial example.
The equality $m(A_n \cap A_k) = m(A_n)m(A_k)$ means, from the probabilitic point of view, that the $A_n$ are pairwise indépendant. Thus, you can try with something like $$ A_n = \bigcup_{0 \leqslant k \leqslant 2^n - 1} \left(\frac{2k}{2^{n + 1}},\frac{2k + 1}{2^{n + 1}}\right) $$ Then show that $m(A_n) = \frac{1}{2}$ for all $n$ and when $n \neq k$, $m(A_n \cap A_k) = \frac{1}{4}$.