Doubt in neural network units

[Advanced Learning Algorithms] > [Week 1] >

How there are different w1, b1, w2, b2, w3, b3? As we will be passing same a[2] to all the same three activation functions (sigmoid)? Can someone please explain?
Thank you

Because we setup layer 3 to have three nodes. Each node has one w and one b, so layer 3 have three pairs of w and b. Each node has its own pair of w and b in order to learn different things from the same a[2].