Jared Bernstein has a chart of the Federal Reserve's steadily falling forecasts for real growth in 2012.
Now here's an important point to note about this. During this time, the Fed hasn't been revising its inflation estimates upward.
Why does that matter? Because there are two reasons why real output could come in at a lower level than was previously anticipated. One would be that the economy proved less efficient at translating demand into real output than you expected. In that case what you'd get is less real output than you expected but more inflation. Another would be that the level of demand in the economy proved to be less than you expected. In that case what you'd get is less real output than you expected, but the same or lower inflation. We've been in Scenario 2—less demand than expected.
But when it comes to the Federal Reserve, how does that happen? The Fed, after all, isn't a third-party forecaster. It's a policymaking institution. In particular, it's a policymaking institution that drives the level of demand in the economy. So when they revise their "predictions" of future demand downward, they're not just offering a guestimate, they're making policy and shaping expectations. They could have said, "We're committed to 6 percent annual demand growth, and our estimates of the output/inflation tradeoff will vary," but instead they said, "We're commited to sub-2 percent inflation, so total demand will be as low as it has to be to hit that target." Doing that has required steady downward revisions to real output.