I see from this paper that compared with Way et al. (2022), they introduced "learning in operational costs, rather than only in CAPEX", which benefits solar and offshore wind, also "solar power and wind energy see a higher learning rate than previous model versions". So very surprised wind not gaining more.

It is difficult to compare the results of Way et al. (2022) and this paper directly since in the former final and usable energy were reported and here it is electricity that is reported in the text, although from their relative share (both across time and wind vs solar at a given time), the conditions for solar is probably more favorable and wind growth is more constrained in this paper.

(Note: if I recalled correctly, Way et al. were the first to develop this system dynamics model that made learning rates endogenous feedback processes.)