True or False: Increasing the depth of a neural network always leads to improved performance, as deeper networks can capture more complex features without any drawbacks.

False. Adding more layers can make a network able to learn more complicated patterns, but it also creates problems such as vanishing gradients, overfitting, and increased training time, so deeper does not guarantee better results. When a network gets too deep, the error signal can become too weak for the early layers to learn, causing the model to degrade in accuracy instead of improve. Regular techniques like skip connections or batch normalization help mitigate this, but the depth alone is not a silver bullet. For example, a 10‑layer network might outperform a 3‑layer one on a simple task, but a 100‑layer version without proper tricks may actually perform worse because the gradients vanish and the model overfits.