Rob Freitas has a good report analyzing how to use molecular nanotechnology to return atmospheric CO2 levels to pre-industrial levels by about 2060 or 2070.
My only complaint is that his attempt to estimate the equivalent of Moore’s Law for photovoltaics looks too optimistic, as it puts too much weight on the 2006-2008 trend, which was influenced by an abnormal rise in energy prices. If the y-axis on that graph were logarithmic instead of linear, it would be easier to visualize the lower long-term trend.
(HT Brian Wang).