my $p = sub {
    my $n = 2;
    my $g = sub { $n++ };
    sub {
        my $r = $g->();
        $g = sub {
            my ($f, $g) = @_;
            sub {
                while (1) {
                    my $r = $g->();
                    return $r if $r % $f;
                }
            };
          }->($r, $g);
        $r;
    };
}->();

while ((my $n = $p->()) < 500) {
    print "$n\n";
}
[download]

It's a prime number generator. It starts with an iterator

$g = sub { $n++ };
[download]

which is captured by the closure $p.
each time you call $p it gets the next value $r from $g then replaces $g with a filter that calls the previous $g until it gets a value that is not a multiple of $r; finally $p returns $r.
It's a technique called the "sieve of Eristothenes": having got a prime number eliminate all multiples of it from your input, then the next number you get must also be prime.

perl -e 'print sub { "Hello @{[shift]}!\n" }->("World")'
[download]