I've read the paper, and can assure you there is no breakthrough in there.
It's just phrasing primes and related functions in terms of a wheel sieve with basis {2,3,5,7} [1].
[1] https://en.wikipedia.org/wiki/Wheel_factorization
The internet as we know it would quite literally break if someone were to discover any sort of function that can return the first N primes, no matter what kind of function it may be.
However I hardly understand how the wheel factorization works at all, though. How is deterministically assigning prime numbers to four buckets going to solve the task?