Can Uniform random variables be viewed as "more random" than Normal random variables? Since a Normal random variable has a mean which has a higher likelihood of observation compared to its tails, would it be correct to say that a Uniform random variable (where mean and tails share the same likelihood of observation) can be viewed as "more random"?
My reasoning is that since a Normally distributed random variable will produce observations with higher likelihood of observing certain numbers (e.g. the mean), there is thus a higher element of predictability and thereby lower complexity [1] in its observations compared to observations from a Uniform distribution.
Edit: by "more random" I was referring to an increased (Kolmogorov/Solomonoff) complexity in the second paragraph [1]. I should have been more explicit.
[1] https://en.wikipedia.org/wiki/Kolmogorov_complexity
 A: It depends on how you define “more random”. Obviously, you would need to compare distributions of a similar scale. Otherwise, for a given uniform distribution $U_{[-d,d]}$, one can construct a normal distribution $N_{0,100d^2}$, which will look “more random”.
Two standard ways to fix “the scale” of distribution are 1) fix the finite support 2) fix the variance. The standard metrics of continuous distribution randomness is differential entropy.

*

*What distribution has maximum differential entropy given a support $[a,b]$? It's a uniform distribution $U_{[a,b]}$. However, normal distribution cannot compete here, since it always has infinite support.

*What distribution has maximum differential entropy given a fixed variance $\sigma^2$? It's a normal distribution $N_{\mu, \sigma^2}$.

The way to intuitively explain why normal distribution beats uniform, is that tails of uniform distribution are not uniform. They are zero. If you give me a variance of uniform distribution, I can guarantee that the random value won't be larger than this and won't be smaller than that. And this is what limits randomness a lot.
