Of course, digitizing everything was the goal. And considering what Google has in the bank, $125M is cheap to settle. The bad news is that it's a settlement, and doesn't make any difference in the law, but it does set a non-legal precedent.
Of course, I'm curious if OCLC's fledgling copyright information registry is partnering with Google; it would make sense not to duplicate effort, and crowdsourcing has made sense in other fields. It certainly hasn't hurt LC, with its Flickr experiment.
I'm not as concerned about Google replacing libraries. It's not like Google is taking the books away from libraries, stopping ILL, or burning them. It certainly has a lower barrier to entry than the aggregators, when an individual can pay for full access to only what they want. Certainly better than Corbis, who bought the photos and has locked them up.
That's not to say that it can't change, but the libraries have gotten their digital copies, presumably the Internet Archive will crawl the open collections, and the libraries will know what was actually used if it comes to them doing the work again.
What's the worst case scenario? Google goes out of business, all the digital files go away, and the libraries are no worse than before, and some publishers and authors have made some money they wouldn't have gotten under the first sale provision.
What's the upside? 20% instead of snippets, payment to authors (if they can be found), fewer dead trees, universal access.
I don't have any problems with that.