I wonder if the author had some more profound message?
Is it about pulverizing software artifacts into token statistics?
What remains is not "code", but code-shaped matter?