I don't understand what the fuss is about except that its trendy to dump on Facebook for any and every conceivable or even illogical reason. All cyber technology comes with and implicit and inherent danger that it will be either used nefariously or misused to harm. It was and remains perfectly legitimate subject matter. Kurzweil in his discussion of the Singularity has made similar points in his books and you don't have to believe in the Singularity to understand the technological inertia he cites.

Most recently, AI folks discuss the same subject matter; https://www.edge.org/conversat...

If a design attempts to answer the wrong question, no manifestation of that design can correct the original flaw.

The issue of gathering requirements is complex. One must distinguish between the raw and the cooked. Is the software breaking new ground or a refactored design of an existing system?

Care needs to be taken that legacy redesigns re-examine the premises that the original systems tried to solve.

a.) What was the system trying to accomplish and is that relevant today?

b.) Are the technological constraints that applied to the manifestation of the legacy system obsolete? today's solution to the original problem may require not a refactoring but a retooling as well (architectural considerations).

The design (including requirements) is the most important artifact of all processes. The design instructs all subsequent implementations regardless of components or coding conventions du jour.

Design also must define and inform cybernetic context. Where's the by-product fit? Where's the architectural edges? And what is it overlapping? In other words, how can a potential user or consumer make sense of the thing in plain verbal and visual language?

The input to design is the question to answer. The output are precise algorithms a tool to accomplish the task will need to accurately conform to.

How the design is built, imo, is out-of-scope and quite frankly an irrelevant distraction.