Despite classic logics denial of authority, once you subordinate logic to observation (Bayes) then authority is actually a good positive indicator. Lack of authority however is a poor negative indicator. Long story short, authority only adds support indirectly, but that's useful more often than many would like to admit.
I think you're simplifying this into a pro/con dilemma. Stephenson is obviously very pro internet. So what he's pointing out is really a lack of ambition in using the new computational and communication toolkit we've built. Digital economies are fundamentally different, despite the bubbles using this to hype absurdly bad traditional business plans, this basic characteristic remains. We have not even begun to explore the implications of that.
We're too busy gratifying the desires born of an older world.