I dont think anyone who seriously understood the subject at the time believed that chess engines as a concept could not make novel strategies. They could believe that the current(for their time) implementations of a chess engine could not, but understand its inevitable that it could. Same as the fact that we dont believe chat gpt can invent or create anything truly novel, but we know that AI and neural networks as a whole can and almost certainly will. As to whether or not terrorists can use it to make weapons, they probably can benefit from it the same way they could benefit from a much improved version of google. In that there can definitely benefit, but not completely handle the process.