There's also the flip side to this, "doing too little work before looping in others". Management brings overhead. Meetings bring overhead. Not every (sub)project needs multiple ICs.
That said, I think the E2E MVP/"tracer bullet" approach is a good compromise. It sets a clear milestone, which could be achieved by a single IC, but which provides something tangible for a team effort to build on (or meets intractable obstacles and and dies prematurely, saving everyone time before the PM engine is spun up).
That said, I think the E2E MVP/"tracer bullet" approach is a good compromise. It sets a clear milestone, which could be achieved by a single IC, but which provides something tangible for a team effort to build on (or meets intractable obstacles and and dies prematurely, saving everyone time before the PM engine is spun up).