How to handle error with gradient descent fitting

Question

I have a set of points x and y:

sport numpy as np
np.random.seed(11)
x = np.linspace(0, 10, 98)
y = 5.0*x + 2.0 + np.random.uniform(-1.0, 1.0, x.shape)