The new information is in the distribution parameters that were updated based on the Mt.Gox data.
The Mt.Gox leak was the data after accounts were frozen right? Considering there were many hints of Mt.Gox's impending failure well in advance (goxing was a verb a long time ago), I expect a large percentage of people withdrew their funds either completely or mostly (my case). Therefore, I am not sure how useful raw Mt.Gox numbers would be for finding the overall distribution.
20 pages back in this thread, when the methodology was created, we used the distribution data from a certain company called Silvervault as a basis for some assumptions of this model. Even impaired Mt.Gox data is an improvement. An example problem where we have better answer now, is:
1.0 million coins belong to people with 100-1000 coins
There are 45,000 people with 10-100 coins.
How many people own 100-1000 coins?