> I think the observation that we need a way to permissionly interact with data is spot-on

The entire argument rests upon this premise and the author doesn’t explain it one bit.