If the question was about constant time complexity but the questioners had a secret additional requirement that they refused to share, that's a flaw in the questioning. If the questioners had that additional secret requirement and understood amortized analysis, then they should have at least given a hint at their secret requirement by saying something like "your proposal does indeed deliver amortized constant time complexity, but what could the potential downsides be of this approach?" If the commenter is giving an accurate description of what happened in the interview, then the questioners were absolutely the ones being stubborn or ignorant. (And, incidentally, I'm pretty sure that dynamic arrays are going to come out ahead in practice even with the secret "tight loop" requirement.)