I disagree with this. Often what is gained from a proof is not the fact that a proof is known, but more insights about the original problem. So, a small technical error may "invalidate" a proof but it does not make it meaningless. Just like a bug in software does not make it worthless.