Indeed I don't believe you don't understand. But, just in case:

$foo->{bar}->{baz}->{quux};
$foo->{bar}->{baz}->{quux}->@*;

${ ${ ${$foo}{bar} }{baz} }{quux};
@{ ${ ${ ${$foo}{bar} }{baz} }{quux} };
[download]

I don't see why ->@* in the left-to-right version doesn't enhance readability just as much as the previous part of the expression, compared to the inside-out version. I must also say that writing the inside-out version was difficult for me, and reading it, even more so. If $foo->{bar} is better then ${ $foo }{bar}, then I don't see why $foo->@* isn't equally better then @{ $foo }. You feel differently, so how about you explain why circumfix is worse in some situations, but becomes better 'at the top level'. "I can tell at a glance" isn't a good explanation because slices exist, e.g.:  for ( @{ $foo->{bar} }{"a", "b", "c", "e"} ) .... You can't even tell "at a glance" you're dealing with a hash!