0z gfs run rolling in so far it looks like it will miss the Caribbean to the north on this run, however Euro will still likely take it into the Caribbean. Closest approach for the Caribbean is Thursday night into Friday. At this far into the run (+186 hour) it's slightly south of the prior 18z run. Ridging to the north is fairly strong also.

After this, it gets very close to the Turks and Caicos islands in the Bahamas, enough to be a problem there. Into the central Bahamas late August 12th into the 13th. Bermuda high is strong, major hurricane over Nassau on the 13th.

Late evening on the 13th, its very close to Florida, very close or just over Cape Canaveral (925mb major hurricane) but stays just offshore (very similar to Matthew), eventually landfalls near Savannah/Hilton Head on Aug 15th, a bit weaker, then rides inland through North Carolina.

earlier runs of the CMC misses the islands and takes it out to sea, the UKMET rides it through the northern Caribbean islands (weak)

The newer 0z run of the CMC is weaker, but much further west.

Splitting the difference takes it very close to the northern Caribbean islands, maybe skimming Hispaniola after missing Puerto Rico barely to the north.

HMON/HWRF develop it into a hurricane in about 5 days.

In short still too soon to tell impacts, about a 50/50 split for Caribbean impact, and much less for US/Bahamas impact, but more than earlier. A slight shift to the west. The most alarming thing about the GFS is the 600dm ridge that holds for several days, which would likely force the storm west and raises the chances of land impacts a great deal (taken in isolation, if the trend holds, even more so)

99L needs to be watched VERY closely over the next week and a half.