Now you're changing your story. PRISM was widely reported not to provide automated/unauthorized access to data; it is now believed to be a submission/management system for FISA requests, which are reviewed by providers before the providers own personnel locate the data. The Guardian silently (to their shame) retracted the claims that PRISM was anything more than that.

Meanwhile, illegal direct access to intra-provider communications can be cured by improving operational security at providers, which is exactly what the story we're commenting on is an instance of.

No, I don't think your argument is valid.