Popper’s bridge - and why it repeats the induction fault

To oversimplify it: according to Popper, induction is not an acceptable scientific method, since a finite number of observations is unable to verify the truth of a statement. It is always possible that we would find a counterexample in the future. So the solution is falsification, and passing more and more observational tests is not a cause to increase our confidence in a hypothesis.
Now imagine that we have to build a bridge and we have two physical theories: an older one which is “tested many times and has passed every test… and a brand new theory that… has never been tested." Applying Popper’s logic, there is no an essential difference between them: neither of them falsified, so they are equally “reliable.” 
 Obviously, it would be simply stupidity to choose the newer hypothesis, although Popper didn’t have a good answer for this contradiction [Godfrey-Smith, Peter: Theory and Reality 2003, p. 67 - 68].
Popper introduced his concept to bypass the problems of induction. The bridge problem pointed out the weakness of a theory based solely on falsification. It is a similarly serious counter argument that Popper’s falsification concept based on induction, although his aim was to eliminate the logical problems of it.
He stated that a counterexample can disprove a theory. In other worlds: the result of a sole observation can be extended to every future incidence. But this presumption based on an induction, namely that if we repeat the observation than it will necessarily disprove our theory again. But following Popper’s logic which refuses the validity of induction, falsification can prove only that a model/concept/etc. actually don’t work, but cannot prove that it won’t work in the future – after all, Popper’s central idea was the rejection of induction.