Irma is looking a little more ragged right now due to a little bit of dry air intrusion and it is moving WSW, it's a bit south of the 0z EPS model guidance so shifts may occur. I think the models are overdoing intensity in the long range, next up are the Caribbean islands which will likely see some of Irma (how much depends on the exact track, we could get watches starting late today or tomorrow)

Conditions are expected to get better for development east of the Bahamas, where the official forecast takes it very close to the Turks and Caicos at the end of the 5 day period, at this rate parts of the US may be in the cone starting Late Tuesday or Wednesday. Direct affects wouldn't likely be until Sunday or Monday (10-11thh), which means any final preps probably Friday or Saturday depending on where it goes.

Beyond 5 days it is still too soon to tell, the forecast lounge has more speculation and odds.