Recent studies warn that the U.S. economy may return to a phase of secular stagnation. In the next 20 to 50 years, U.S. economic growth will be negatively affected by lower contributions of hours worked and education. But some studies also add that productivity could decelerate sharply and that GDP per capita, by focusing on the average household, neglects that income has already been stagnating in the last 30 years for the households in the bottom 99% of the income distribution. After reviewing recent long-run projections, we argue that similar warnings were issued in the past after all deep recessions. Interestingly, pessimistic predictions turned out to be wrong neither because they were built on erroneous theories or data, nor because they failed to predict the discovery of new technologies, but because they underestimated the potential of the technologies that already existed. These findings suggest that today we should not make the same mistake and undervalue the effects of the information technology. Finally, we discuss a number of issues that should be tackled by future research.