Projections of the gross world product GWP for the 21st century are computed on a simple climatendash;macroeconomic model using different global mean surface air temperature projections provided by General Circulation Models GCMs as input data. Two alternative specifications of climate damage functions proposed by Nordhaus and Weitzman are considered. High uncertainty of long-term global macroeconomic dynamics with respect to the choice of climate scenarios and climate damage functions is revealed. Strong nonlinearity of the Weitzman function combined with the "worst-case" temperature scenario yields a very dramatic scenario of long-term global economic development. A high degree of uncertainty accompanying existing assessments of climate#x2013;socioeconomic projections urgently calls for more detailed and better justified estimations of anticipated climate damages at high temperature increases above pre-industrial level.