The question I see here is whether you are interested in Mathematics as end, or in the Application of Mathematics as an end. Certainly, if all you need is an algorithm that works, don't write a formal proof. However, if your mathematical work is foundational for some future algorithm, you need to know that p(x) holds for every x, not just a bunch that you tested.
Most students do not need to be concerned with (and indeed, are not instructed in) formal proofs. But, if we are going to push math forward, we need to lay a solid foundation of axioms and proofs.