News flash: every mainstream economist today, including Bush's cheif economic advisor, accepts the Keynesian theory as the best explanation of short economic fluctuations. Samuelson is no exception. The empirical evidence in favor of it is overwhelming. You need to study.
And Keynes was not a socialist.
And no, we are not all Keynsians now. Orthodox economics was upheld by the Austrians throughout the Keynesian episode, and most Keynsians themselves accepted the monetarist criticisms of their position formulated in the 1970s. (When Keynsian prediction was so systematically wrong for so long, only religious ardor could keep anyone believing in it). Monetarism is a half way house - it upholds a few of the theorems of orthodox economics. Keynsians who do not accept even those are so obviously just wrong that no one takes them seriously anymore (nobody believes in government spending "multipliers", for example). Only those on the left still believe a word of it. As a theoretical position, it was demolished long ago. Empirical econometrics, on the other hand, these days pays as little attention as possible to economic theory.
But that's the key here, short-term. The spending multiplier effect still creates spending for the sake of stimulating short-term employment. Later, the piper must be paid by either raising taxes or running budget deficits. There are better ways to stimulate employment besides digging holes that you'll just have to fill at a later date.
"Keynes was not a socialist". He was not, yet his solutions for economic problems led to socialism.