Jason's sweeping "it's less than <1%" comments are getting tired.
We can't tell whether Jason is misleading us about the proportion of scrape-generated pages on Mahalo without access to any Mahalo page statistics.
I'm not for or against Jason on this matter, I'm just saying that we have no data on which to base any conclusions. It's possible he's telling the truth.
It would be interesting for someone to take up the challenge of creating a small web app that finds all Mahalo URLs, heuristically examines them for spamminess and generates some statistics.
Not that such an app would be of any particular long term use, but it might be interesting nonetheless.
Mahalo is a great money printing machine, but come on, those are pages designed to do one thing - rip content quickly, monetize, and SPAM google.