Other than JS performance being downgraded when the app is built this way. It appears that the actual facebook servers that serve the customised content are also slower (read: much slower.)

The speed experienced can't be attributed to JS alone, there simply isn't enough JS being processed when many of the 'slow' encounters are being felt, additionally the speed fluctuates through the day/week which would be consistent with server load rather than a small JS not loading and running quickly enough.