> They cannot claim a copyright to exhaustive computer generated corpus.

And why not? For example, a UK law specifically mentions "computer-generated" work:

> If the work is computer-generated, copyright expires fifty years after the work is made.

https://en.wikipedia.org/wiki/Copyright,_Designs_and_Patents...

Is there any law you based the word "cannot" on?

Is there a limitation on the extent of such copyright claims a person can make? Even though physically one couldn't write it on paper during one's lifetime?

In the US, you cannot claim copyright on any work devoid of any creative intent. A phone book is not copyrightable, but the specific fonts and organization can make a given phone book copyrightable. Organized trivia isn't copyrightable, but the expression is. There's no expression with an extensive corpus.

It also wouldn't matter. The existence of the corpus doesn't change the copyrightability of other works. You'd still have to convince a jury it was more likely you were inspired by the corpus than the hit song whose melody is identical.

This isn’t reading the dictionary, it’s similar to generating new unique sentences and claiming copyright on that. Right?

It's creating 12 long sequences from 12 notes. That sounds far closer to a dictionary than anything else.