Yes the sun does not shine at night indeed and there are some losses in converting sun rays into a usable energy using regular solar panels. But exactly the same issues are present in the proposed co2+h2o reactor project.

This does not make any sense. A simple chemical reaction of methane burning is CH4 + 2*O2 -> CO2 + 2*H2O + 890 kJ/mol energy freed during the process.. So in order to convert CO2 and water back to methane and oxygen you need to spend the same 890 kJ/mol energy. Okay, you get the energy from the sun, but why not to use the sun to generate the energy directly without doing the CO2 conversion back and forth?

I haven't noticed any improvements yet. The boot times are painfully slow just as before SP1.