Since we sample two cameras, the random variable X can only have three possible values:
0 - both cameras work correctly
1 - only one defective camera
2 - both cameras are defective.
To fully define the random variable, we need to compute the proability for each of the possible values.
X follows a Bernoulli distribution, where we want to count the number of "successes" (finding a defective camera) on a sample of 6 elements. Since 4 out of the 6 cameras are defective, the probability of finding a defective camera is 4/6=2/3.
In general, we have

So, the probabilities are


