I think you are misunderstanding the situation. Given code like:

    if (!p) {
        use(*p);
    }

(given no previous knowledge about p) no compiler will remove the "if (!p)" part.

What people are complaining about is the opposite case:

    use(*p);

    /* The compiler reasons that if p == NULL, the program would have crashed by now,
       so if we got here, p != NULL must hold. */

    if (!p) {  // the compiler can remove this branch
        report_error();
    }