Suppose in mathematics, someone finds a completely new elementary proof of a famous theorem and wants his work ''peer reviewed'' on a site like MathOverflow or MathStackExchange before submitting to a formal journal. How safe would this be ?