4 ms·
≈ is a loose approximation. The fact is that `e * (n!)^(1/n)` is quite close to `n`. But John didn't factor out the `e`, causing the confusion in this HN pos
by gowld 3y ago
≈ is a loose approximation.
The fact is that `e * (n!)^(1/n)` is quite close to `n`.
But John didn't factor out the `e`, causing the confusion in this HN post's threads.