Kind of a terrible title for the article, IMO.
Google basically caved on the "opt-out." Rights holders have to explicitly give permission first.
Most of the titles in question are obscure out-of-print or orphaned books, with low-quality OCR scans, whose primary constituencies will be academics.
Probably not big sellers.
If it does take off, Amazon could always engage in their own scanning project, presumably on the same terms, if they were willing to incur the costs.
Still, in general it's a positive step, especially if the US takes the cue and switches to opt-in.