This is a step in the right direction - the right direction being towards a decentralized peer-review system which cuts out the middle-man!
I think that jumping directly from the centralized publisher filter system into an open and decentralized system is too radical and scary for the establishment. Most wouldn't want to 'waste' a good paper by putting it through such an experimental system before it has proven at least some reliability. So having a reputable publisher take this transitory step is a very positive development.