Ignoring the ethics of downloading the images from comics that rely on advertising to stay in business (wouldn't it be better to mirror the whole page?), here are a few pointers on your code.
You might want to test to see if get got anything. If it returns undef due to failure your script will break. For this reason it's usually best not to use LWP::Simple
If you use m// with one of the alternative delimiters (@, #, | etc) you can can save yourself escaping backslashes and having code that looks like line noise
On the same note, look into q// and qq// for quoting
/.*\.(...)/ is silly. It would make more sense to use something like HTML::LinkExtor to go through the links and find one with the filename/image dimensions you want. Parsing HTML with regex's is generally not the right way to go
You want either open PAGE, ">comics.html" or die $! or open(PAGE, ">comics.html") || die $!
Perl has a foreach loops which saves you from using error prone C style for loops.