Long range speculation includes watching the global (and other) forecast models to watch for trends, when a system is still an invest and not fully developed things can (and do) change wildly from run to run, but they tend to vary less as time goes on. These models are helpful in that they give a general path to look for, and we look for consensus and stability between runs to gauge confidence in what they do. many times the position of the system in the model isn't as important as the conditions (Ridges/troughs/lows/upper level lows/dry conditions) around it.

That said, here's the first model set for 99L:

0z Euro: Not as strong as it was on the prior run, but it has it crossing the leewards as a Tropical Storm Tuesday night, then falling apart int the Caribbean afterwars (another system near the Yucatan starts going in the Bay of Campeche).

6z GFS: Develops a hurricane around 40W in the Central Atlantic Monday morning, keeps it north of the Caribbean islands, and bends it back west (north of the Bahamas), and has a major hurricane landfalling in coastal Georgia near Brunswick/St..Simons on Thursday August 17th (2 weeks from today). The 0z run prior to this had the same result. After landfall the 6z GFS turns it northward inland over North Georgia then eastern Tennessee (Knoxville) into Kentucky and Indiana weakening rapidly. (with a ton of rain).

The fact the models want to push this so far west implies the high will be strong around that time, which would increase the odds for some sort of impact on the land US down the road, if it doesn't stay south and go through the Caribbean like the current Euro Shows. So some sort of impact is likely, to what degree and where (islands vs US) is still up in the air. If the trend moves east during the next few runs, out to sea becomes more likely.

Way too early to say which is going to be more correct, nothing has developed yet, and the first runs are typically very wrong.