A simple way to generate random points on a sphere is by generating random points in a cube until one falls inside the sphere, then normalizing it to push it to the sphere's surface.
To generate random points on the unit sphere, independent random values are generated on [-1, 1], and if the sum of their squares is less than or equal to 1, the normalized values are returned.
This approach is intuitive, with minimal dependencies compared to the more common method of generating n independent values from a standard normal distribution then normalizing.
While the accept-reject methods used in this approach may have varying runtimes, they are practical for most cases. The efficiency trade-offs and considerations in parallel computing and cryptography are also discussed.