Paul Krugman Gets It Almost Right
The one thing I would object to in his post is the fact that he only draws a distinction between short- and long-term effect on demand, while assuming a completely inelastic supply, or in other words a vertical supply curve. While that may be if not completely true, then at least approximately true in the short run, it is simply not true at all in the long term. Although Krugman's favorite politicians in the Democratic Party greatly limit through environmentalist regulations the possible domestic supply increase, the supply response in less environmentalist countries like Brazil is a lot stronger.
Moreover, the non-vertical nature of the long-term supply curve actually does not depend on whether we've reached "peak oil" or not. Even if that was actually true, and it would be impossible to increase supply compared to current levels, it would nevertheless be the case that higher oil prices would raise supply compared to what it would have been at lower prices. This is to say, if prices were lower then supply would have fallen faster, instead of being unchanged or falling slower. Which in terms of Krugman's diagram would have implied a upward sloping long-term supply curve, with a "P3" price below and to the right (implying a higher quantity) of P2.