These articles are disappointing. First, people have already long since explored this design space for things like network routing and holding it up like it's a new idea is academically fraudulent. (I don't know whether it's the journalist or the scientists without more research, but I'd lay money on the journalist.) Second, the real world inspiration of ants is somewhat limited by the fact that the ants have been thrown at problems where, in computational terms, they have a lot more ants than complexities in the problem. A naive translation of the "ant algorithms" to computer programs would produce a very, very, very slow solution.

It's sort of cool, it's a nifty technique with limited applicability, but can we please stop acting like antlike-automata is some sort of cutting-edge brilliant breakthrough rather than a relatively standard tool in the belt for some domains? It's not the next generation and arguably not even the "current" generation.