It seems you missed the conclusion in which they state that there is unlikely to be a learning curve beyond the build of the first unit or two as evidenced by F

It seems you missed the conclusion in which they state that there is unlikely to be a learning curve beyond the build of the first unit or two as evidenced by France and Japan's build times. They also did a decent job of showing that the average build time should be around 4 years which is equivalent to the build of a coal power plant and that the time it takes in both cases is determined by how long it takes to build the boiler and turbine. The swelling of timelines appears only when large changes in regulation enter the picture.