This is a pretty poor article for the following reasons:

• The semantics the author is describing is roughly S3, which turned 10 last year. So not really a “new paradigm”.
• Object stores are are even older, what do you think people stored their data in before RDBMSs.
• The semantics of an Object Store are not really concrete. S3 has a hierarchy, for example.

And the part that made me say WTF out loud:

File systems aren’t made for HTTP(S) but humans. Sharing and managing files in a file system is difficult to handle programmatically (think of the cunning C/C++ file handling tricks that most of us struggled to understand well). Handling file streams and the possible boundary cases is error prone and takes lot of time and effort.

Sockets, HTTP protocol, another computer!, all of these make using a file system “less difficult”? I mean, it’s unclear what “sharing and managing” means, but file systems have great programmatic interfaces at this point, they freaking have to.