I completely agree with you.

However, he did 'break into' an MIT switch closet to run 'keepgrabbing.py' over a 1Gbit/s connection. He wasn't just downloading 100KiB PDFs, either. He downloaded at least two million documents. The indictment isn't clear exactly how many, and it sounds like he downloaded a lot more than 2 million, to boot. Not all of JSTOR's documents are neat 100KiB PDFs, either: a substantial portion are scanned images (1+ MiB PDFs) from old journals. So, we're looking at the TB range of data.

This is not to say that his intentions were ignoble...