Let $A \in \mathbb{R}^{n \times n}$ be a positive semi-definite (PSD) matrix, and $B \in \mathbb{R}^{n \times n}$ a diagonal matrix where all diagonal entries are between 0 and 1. Is it true that $A - BAB$ is PSD? That is, for any vector $v$, that
$$v^T(A - BAB)v\geq 0$$
The best way to address those problems is to look for a counterexample. Take for instance,
$$A=\dfrac{1}{10}\begin{bmatrix}6 & 5\\4 & 4\end{bmatrix},\ B=\dfrac{1}{10}\begin{bmatrix}4 & 0\\0 & 8\end{bmatrix}.$$
The matrix $A$ has eigenvalues $0.9583$ and $0.0417$. However, the matrix
$$A-BAB=\begin{bmatrix}0.504 & 0.340\\ 0.272 & 0.144\end{bmatrix}$$
has eigenvalues $0.6774$ and $-0.0294$.