It looks to me that it may be nearing time to toss out the model runs on Karen of the past several days. As of this morning, Karen has all but stalled, or has even been tracking a little bit to the southwest, if I buy the Extrap and my own eyes. Best I can tell, the LLC hasn't gone much of anywhere during the eclipse, and now, at the precipice of 50W, Karen is not yet even north of 14.5N, which was the approximate leftmost outlier for her position by this longitude, from the model suite average of the past several days.

This is the official word from NHC at 5AM
0900 UTC FRI SEP 28 2007
TROPICAL STORM CENTER LOCATED NEAR 14.1N 49.8W AT 28/0900Z
POSITION ACCURATE WITHIN 30 NM

PRESENT MOVEMENT TOWARD THE WEST-NORTHWEST OR 290 DEGREES AT 9 KT

I just don't see it. I might be able to believe that is her motion of the last one hour, or so, but certainly not the average of the past five or six hours. In fact, this take is entirely substantiated by comparing the 11PM advisory position with today's 5AM.

11PM last night:
INITIAL 28/0300Z 14.8N 49.5W 50 KT

So, Karen went from 14.8 N 49.5W to 14.1N 49.8W Averaged out, that reads a lot more like SW, or even really SSW, over 6 hours. Karen is a highly-sheared tropical cyclone with intermittent bursts of deep convection within and about her center. However, this convection does not appear to have been consistent enough to allow her to follow the marching orders of all of these models.

Certainly this may be a relocation on the assumption that we didn't know where her center really was six hours ago, but looking at the night vision loops and microwave passes, that initial fix at 14.8N 49.5W looks to have been pretty darn good. I think Karen just went .. left.

With shear forecast to slacken some in the 2-4 day out time frame, and Karen likely to often remain rather shallow until then, and already having strayed considerably left of consensus, a consensus among models that now nearly unanimously send her westbound on Days 4 and/or 5, as it is.. I do think it is very close to, if not already, time to rethink these model forecasts.

This last comment from Mainelli from the 5AM Discussion really sums Karen up, methinks:
IT SHOULD BE NOTED THAT THERE ARE LARGE UNCERTAINTIES IN BOTH THE
TRACK AND INTENSITY AT THE LONGER FORECAST PERIODS.