Could someone explain why validator selection being pseudorandom instead of truly random is not a risk to ethereum.

Proof of stake ethereum uses RANDAO where the 32 validators in the epoch each provide a number that get XOR (bitwise computer operation) together that then decides the the n+1 validator set. Because it’s just code it’s at best pseudorandom and if the 32 validators work together they can heavily influence who the next validators are. Pseudorandom number generators vs truly random number generators( PRNG vs TRNG) is a topic that has been studied in cryptography for a long time and it is known that software alone cannot be truly random. Isn’t this a security concern? I believe that there are plans implement hardware devices for “verifiable delay functions” which just introduces mining hardware back into the protocol. The whole point of proof of stake is that you don’t require hardware running that’s just using electricity. It’s essentially using a variant of proof of work to seed the randomness.