Basically he's right, as I've argued in my book Risk and Business Cycles. Here's a bit of what he is serving up:
What happens, instead – or at least that’s how I read it – is that Austrians slip Keynesianism in through the back door. Implicitly, they associate booms and slumps with rising or falling aggregate demand – utterly unaware that their own theory doesn’t actually make room for such a thing as aggregate demand to exist, or at least to affect overall employment. So Austrians are basically Keynesians in denial – self-hating Keynesians? – pretending to themselves that they’re not using ideas that are in fact essential to their story.
Sraffa first made a related point in 1932, though without reference to Keynesianism of course. The strongest defense of the Austrians is something like the following. The simplest IS/LM or AD models are models of flows, not stocks. Arguably the Austrians could be pointing to a longer-run stock condition — concerning capital, savings, and the like — which means that the flows of the boom eventually must be reversed into a bust. The Austrians could (though many don't) buy into Keynes as a good short-run theory while addending these longer-run considerations of sustainability.
Krugman's point is harder to rebut if you ask the simple questions of why Austrians a) start from an assumption of full employment, b) postulate that in a boom capital goods production rises at the expense of consumer goods production, and c) argue that real wages rise during the boom. Those can't all happen together.
A separate question is why investors don't see inflation, get scared, and contract the structure of production immediately, rather than first expanding it. Or why unforeseen inflation (if indeed it is unforeseen) does not significantly lower the real interest rate that is paid ex post on borrowed funds (no Fisher effect!), thus supporting long-term investments. Or why investors so respond to the short-term interest rate but are so oblivious to the information contained in the broader term structure. Or just ask how much investors estimate future consumer demand by looking at interest rates — usually not much at all and so they are not so strongly tricked by monetary influences on intereest rates.
The point is not to throw out the Austrian scenario altogether, but rather to rebuild it with foundations from bubble theories and Keynesian economics, plus modern finance and real business cycle theory.
Addendum: Arnold Kling comments.