I've converged on "it's a bad idea to _start_ a nuclear project in 2026, but the sunk cost fallacy doesn't quite apply, so once significantly invested it should be finished and run to its operational lifetime".

That is not how it works.

First, you for decades pay 5-10 times more for energy to subsidize every single green energy project... and then oopsie woopsie you still buy Russian gas and dependent on the evil coutries.