I think most people would use the publisher’s website first and then resort to scihub, because scihub requires a doi or publisher’s link to get the paper.

I don’t think this causes much concern, even if so, I believe a good amount of blame should still fall on the publishers and academic systems that encourages gatekeeping knowledge. Especially when these knowledges are generated by public money, then the public should rightfully have access to them.

Yes, from the announcement, it seems true.