The "economics" essay underplays two important issues and elides a third. First, learning curves. You are reluctant to let go of "economies of scale", for which you seem to place, ahem, a thumb on the scale against "economies of mass production". What you miss is the dynamic nature of learning. If you produce two reactors of size X/2 rather than 1 reactor of size X, you get twice as much learning, and that learning compounds. C.f. wind turbines.
Second, nuclear has many more serious competitors than it used to. Geothermal and various storage technologies produce market-friendly energy at gigawatt scale, and capital is being deployed to improve them as fast as it can be absorbed.
Third (and related): all economies are political economies, and nuclear has the worst politics of any technology. To their enormous credit, nuclear advocates are finally taking this issue seriously, but they have so much work to do.
Unfortunately, these three issues feed into each other. The happy case—the one I'm rooting for—is that a 5-10 megawatt technology gets so much capital behind it that costs are brought down to "vaguely competitive with offshore wind", and that it becomes an important option with gigawatts deployed annually starting around 2030. Unfortunately, by that point we're going to have GWh-storage paired with solar that is either 1/10th that cost or twice as efficient as currently, or both. The path for nuclear is narrow and treacherous.