> Is this a brother or a cousin of the "sufficiently advanced compiler"? :-)

I believe the claim was that a sufficiently advanced compiler could do a lot of optimization and greatly improve performance. Maybe my claim here will turn out the same.