Any good scientist will try to support his or her views by pointing to the evidence. What is the evidence that large fiscal stimulus packages have worked in the past? DeLong gives us three examples:
 The 2003-2005 housing boom, facilitated by loose monetary policy;
 The 1996-1998 internet boom;
 The post 1982 boom following the easing of monetary policy, the Reagan tax-cuts, and increase in military expenditure.
He goes on to write:
These are just three examples of a general principle: each major business-cycle expansion we have seen has been driven by a leading wave of spending—by some group that became enthusiastic about their prospects and decided to greatly increase its spending. And that pulled employment and production up.
I view this as evidence that DeLong should have his degree in economics revoked.
This is the best he can do? The first two examples have nothing to do with fiscal policy. The suggestion that the 1980s boom would not have occurred absent the Reagan tax cuts and increased military spending is dubious at best. Moreover, he neglects to point to the several cases we know of demonstrating the converse (with Japan being the most notable recent example).
His thesis appears to be that a significant increase in "confidence" is followed by an increase in spending and an increase in production. This much is no doubt true. Whether this "starry eyed" optimism begins in the private sector or the public sector is, in his view, irrelevant. This is almost surely not true. Confidence in the private sector is ultimately based on information that signals the expected productivity of capital investment (these expectations can turn out to be wrong ex post, of course). In short, private sector confidence is a by-product of changing fundamentals; that is, confidence is symptomatic and not causal. Confidence cannot be manufactured out of wishful thinking; which is the approach that fiscal policy appears to based on.
For more sober appraisal, I refer you to Michele Boldrin's take on this; see here.