Lets just deal with this very simple statement. According to the EIA total losses in the US electrical grid is ~ 5-6%.
It is not mathematically possible to recover more than 5-6% by improving transmission since that is the total lost in transmission. If you did perform this miracle you would have created a transmission system that generates electricity from nothing, defying both mathematics and physics. This isn't arguable.
"Transmission and distribution losses in the USA were estimated at 6.6% in 1997[20] and 6.5% in 2007." -Wikipedia
So it's closer to 6-7% - albeit not a large difference from, "5-6%", but possibly significant when calculating annual costs in trillions of dollars.
What is arguably the larger issue is capacity and maximum load, which nobody has addressed so far. The capacity to distribute power with the current system has an upper limit, if I understand correctly. That means regardless of cost issues, a system with a near-term upper limit is a deal breaker for long-term planning, as long as population increases and concurrently capacity demands increase along with regional population.
Superconducting hydrogen grids have a much higher upper limit of load as I understand it, because the limitations on conventional distribution systems are at least partially demarcated by heat, which is not a critical factor in superconducting high temperature superconducting grids.
Also HTSC grids allow transportation of power over long distances so that regions with high load demand may receive immediate relief from regions with low demand. Since brown-outs /black-outs are another real concern of the electricity distribution issue, this could not be easily fixed by simply increasing deployment of existing conventional technology.
Admittedly engineers with expertise in these areas could clarify some aspects of approach to ameliorating the concerns mentioned in the article I posted, (max load and local over-demand, etc.) which would be interesting to learn about.