This is the gist of the whole thing. The conventionally repeated argument is that bigger is better, but I'm not sure if that's true, all things considered. I don't think it is.
I think that mass manufacturing unlocks a cycle of cumulative improvement through rapid iteration, which eventually leads to way lower costs per energy production capacity. However, scaling energy production profitably requires that energy consumption is scaled as well, i.e. there needs to be new applications which consume a lot of energy, and in general people should be encouraged to use as much energy as possible.
True, but many of the costs outside of the containment vessel also seem to scale - in ways inherently disadvantageous to big reactors.