The video sets out a race between two sets of Primes - just like we covered in Chebyshev's Bias.
Primes ≡ 1 (mod 4) (primes that leave a remainder of 1 when divided by 4).
Primes ≡ 3 (mod 4) (primes that leave a remainder of 3 when divided by 4).
The main discovery here is not so much that there is a bias but that 3 (mod) 4 leads the race 99% of the time. The weird thing is that 1 (mod 4) will always have a chance of being in the lead again at any point, however big the numbets get. Something is bringing the numbers back together. (The contention of this whole site - it is the Riemann Zeta Function).