The upper-level features that propelled Irene to the north may not have been present a week ago, but they were 4 or so days ago. As model guidance changes and as the pattern changes, you've got to go with what the storm is doing, not what you originally predicted. The North Carolina landfall never happened because as the upper-low in the central Gulf moved northward, it shifted the bulk of the westerlies further northward, allowing a ridge to build in near and over Irene, keeping it moving slowly for some time. The ridge, however, was (and is) progressive, which isn't too big of a surprise given the westerly flow that did persist in the region and that the Rex block along 40W was far enough east to allow for the ridge to progress eastward (and now the block has completely broken down).
Three days ago, it became even more apparent that the channel for Irene to move through was coming, and lo and behold it did. Remember -- given weakened steering currents, the natural tendency of a storm is to drift towards the north and west. As a storm moves further north, the natural tendency is for it to recurve out to sea as it meets the westerly flow once and for all. Given this factor alone, once the steering currents collapsed with the storm so far away from the coast, the storm was basically destined to move out to sea.
It was a little surprising to see the steering flow evolve like it did early last week, when things looked more hairy for the SE United States. But, as I mentioned in my blog post from Saturday night, minor evolutions in the steering pattern can make a huge difference, as they did with Irene. To say that the way for it to head north wasn't present/visible isn't really correct; to say that it wasn't there about a week ago is more correct. That's why JB draws a lot of heat, fairly or unfairly -- he's stubborn. I'm stubborn at times too -- I haven't given up on TD 10 quite yet -- but you've gotta know when you swallow your pride and your forecast and revise things. That's why the NHC is usually conservative with their forecasts, in case things change, and why they generally do an excellent job (as borne out by the statistics).
That said, let's see what the next 5 days bring. TD 10's remnants are still around (just as a heads-up, I think Daniel's comments were designed to try to quelch the speculation that TD 10 still existed, when in fact it did not), the flareup in the western Caribbean bears watching (though is largely forced by an upper-low at this time), and the model guidance is in pretty good agreement on something substantial coming off of the coast of Africa in 4-5 days. While Irene may thankfully be heading out to sea, there's still plenty worth watching.
Current Tropical Model Output Plots
(or view them on the main page for any active Atlantic storms!)