The problem arises when we try to use induction to project the likelihood of future events.

The well-used example is that of a turkey in the six months leading up to christmas. Each day he is presented with evidence to support and strengthen his theory that human beings are benevolent and interested in his continuing welfare...until one day.

So while several posters who have made the point that science has great pragmatic value, we mustn't lose sight of the fact that some times it does not and in fact can be damaging if wrong conclusions are drawn from evidence.