Or, you could read what I think he thinks. Which may save time.
My summary. Macroeconomics is the study of fluctuations. Since the 1970's there have been three approaches to macro. First, the new-classicals. These guys build models of the macroeconomy where the aggregate properties of the economy come from the decisions made at the household and firm level. All markets are in equilibrium, and the only thing which creates cycles is changes in technology. This approach is associated with Ed Prescott, Robert Lucas, and Thomas Sargent. Second, the new-Keynesians. These guys emphasised the role of market imperfections like asymmetric information in generating cycles. Top chaps in the NK school would be Greg Mankiw and Joseph Stiglitz. Third, the growth theorists. These guys mainly ignored the problem of fluctuations. People like Daron Acemoglu and Charles I. Jones come to mind.
Blanchard argues that modern macroeconomics has had a sort of convergence from these three schools, where the theory of the new-classicals is augmented with imperfections. The problem with this approach is complexity. Now we need large dynamic stochastic general equilibrium models to make predictions about the economy in the short run, which makes understanding the message behind these models difficult.
Blanchard sees a more `open source' future for macroeconomics, where we hack together macro models based on the problem they are built to solve, and estimate them numerically, without worrying about the general equilibrium and social welfare properties of these models.
I'm a bit skeptical about the issue of pluralism in economics (see here), but I think Blanchard's vision of a broader macroeconomic church, if it came about, would be nice to see if it ever happened.